To our knowledge, this was the first study to assess the impact of sepsis on mortality in a Latin American cohort of patients with hematologic malignancy, as well as factors associated with outcomes. We observed a significant impact of sepsis on 30-day and 90-day mortality. The factors associated with 30-day mortality were SOFA score, severe neutropenia, thrombocytopenia, and hyperbilirubinemia, while those at 90 days were SOFA score, thrombocytopenia, and hyperbilirubinemia.
The 30- and 90-day mortality rates from sepsis were 49.5% and 59.3%, respectively, like those in a study involving onco-hematological patients in a developed country [15, 16, 17]. The rates were higher than those reported for the general populations in developed countries, but like that in Brazil [18]. We expected higher rates in our study considering the patient profile and national data on mortality from sepsis.
In our study, the probability of death at 30 days increased 1.3-fold with each point increase in SOFA score, a finding similar to those of other studies in cancer patients, indicating that the SOFA score is useful for assessing the severity of patients with sepsis, including onco-hematological patients [19, 20, 21, 22].
Studies have shown that neutropenia is a risk factor for infection [3, 23, 24]. In our study, we observed that neutropenia was a risk factor associated with 30-day mortality. A study including patients with onco-hematological diseases admitted to the ICU found that six (20%) admissions were due to sepsis. The 28-day mortality rate was higher in patients with acute myeloid leukemia (37.8%) than in patients with other pathologies (32.4%) [19]. The main risk factors for complications were severe neutropenia (≤ 100 neutrophils/mm3) with an expected duration of > 7 days and the presence of comorbidities such as hemodynamic instability, mucositis, and neurological alterations [25].
Hyperbilirubinemia and thrombocytopenia at the time of sepsis diagnosis were positively associated with mortality. Several studies reported an association between mortality rate and the number of organs with dysfunction (> 3) [1, 15, 26, 27]; however, few studies have reported associations between bilirubin levels and these dysfunctions [15, 28]. In our study, the probability of death increased approximately 13-fold in patients with high total bilirubin levels. Thrombocytopenia is a common laboratory finding in onco-hematological patients and, in our study, was an independent predictor of death. These findings reinforce the importance of monitoring laboratory parameters during the follow-up of onco-hematological patients, especially during hospital stays, when clinical changes can be detected earlier.
The key clinical signs at the time of sepsis diagnosis were tachycardia, fever, and tachypnea. Tachypnea was a common clinical sign (63.7%) and was significantly more common in the groups who had progressed to death at 30 and 90 days than in survivors. Therefore, frequent monitoring of vital signs is crucial. Furthermore, the use of validated scores could facilitate the early identification of these signs by the nursing staff, which, in turn, could inform the medical team [29].
Among the febrile neutropenia episodes, there were a higher proportion of documented infections (MDI and CDI). A study involving the same patient population found similar results [28]. The frequency of CDI was higher in patients with febrile neutropenia who died within 30 and 90 days (15 cases). Subgroup analysis indicated that the rate of pneumonia diagnosis was higher (73.3%) in this population, which may explain the higher mortality rate.
In our study, GNB was the most common infectious agent (78.9%), and the rate of multidrug resistance was high (64.4%). These results may corroborate the national epidemiological results from the SENTRY study, which demonstrated the predominance of GNB in BSIs and a progressive increase in the resistance of these microorganisms to carbapenems [30]. In addition, most BSI episodes (65.5%) occurred during neutropenia. These results are concordant with those in recent literature reporting changes in infection epidemiology in patients with febrile neutropenia, including increased GNB, especially those resistant to antimicrobial agents [23, 31, 32, 33, 34].
This resistance profile may limit the efficacy of antimicrobial therapy and, consequently, increase mortality in patients with severe sepsis or septic shock [28]. In our study, MDR GNB infection was not significantly associated with increased mortality, but there was a trend toward a higher rate of death in the group of patients who died within 30 days. However, the univariate analysis showed lower antimicrobial therapy effectiveness in patients who died, which might have been related to the multidrug resistance profile.
Regarding the adoption of the sepsis protocol, several studies have stressed the importance of the quality of care and have reported a relationship between decreased mortality and protocol compliance [18, 35, 36, 37, 38]. In univariate analysis, protocol nonadherence tended to increase 30-day mortality (p = 0.057). These results reinforce the need for accurate diagnosis and management of patients with sepsis, especially in patients with hematologic malignancies.
The main limitation of this study was the small sample size given the specific characteristics of the population (sepsis in patients with hematologic malignancies) and its retrospective analysis of the data. Notwithstanding these limitations, the results showed that the SOFA score was a predictor of mortality from sepsis and organ dysfunction such as hyperbilirubinemia and thrombocytopenia in these patients, indicating the need for laboratory monitoring and follow-up.
Notwithstanding these limitations, to our knowledge, this study is the first to involve a Latin American cohort of onco-hematological patients with sepsis. The results showed that the SOFA score was a predictor of mortality from sepsis and organ dysfunction such as hyperbilirubinemia and thrombocytopenia in these patients, indicating the need for laboratory monitoring and follow-up.