Sepsis is a life-threatening infection, and there is still no definitive conclusion regarding the analysis of prognostic factors of sepsis in older patients [16]. Many scoring systems are available to evaluate the severity of sepsis in patients, such as SIRS, SOFA, qSOFA, and APACHE II scores [17, 18]. However, these scores not only need to monitor the basic vital signs of patients but also need to test many physiological and biochemical indicators. The process is complex and laborious. The study aimed to construct a prediction model that could have an immediate impact on treatment decisions. It is our hope that this model will not only increase working efficiency but will also provide clinicians with more safe and effective treatment options.
Previous studies have shown that death due to sepsis in patients is strongly associated with some biological indicators, such as PCT, hCRP, renal insufficiency, Lac level, and ALB level [19–23]. Based on previous studies and clinical experience, this study included variables that were straightforward to collect and may have an impact on prognosis, combined with the nutritional status of the patients, and obtained an independent risk factor for poor outcomes after analysis. Anaerobic metabolism produces Lac, and the resultant hyperlactataemia often indicates that patients with sepsis have circulatory or respiratory failure. Lac level ≥ 4.0 mmol/l is related to an increase in mortality at 28 days in approximately 30% of patients with acute illness. A prospective cohort study involving 126 patients, who were diagnosed with severe sepsis or septic shock, found that lactic acidosis can more accurately predict hospital mortality in patients [24]. Therefore, Lac levels can be considered a good predictor of mortality risk with critical illness. Our results matched that of former studies. Therefore, dynamic monitoring of the blood lactate level was significant in evaluating the severity of illness and determining prognosis. BUN is the metabolic outcome of protein metabolism and is excreted by the kidney. When protein is degenerated or glomerular filtration rate decreases, the BUN level will increase. Therefore, the level of BUN can reflect protein catabolism in the body and is also a sign of kidney damage [25]. At the same time, BUN is one of the common indicators of kidney function in the clinical laboratory. Patients with sepsis have significantly increased protein catabolism and tend to have acute kidney injury. All these factors can lead to an increase in the BUN level. Another study showed that the BUN level was an independent predictor of mortality in patients with sepsis [26]. This study also showed that mortality at 30 days increased gradually in older people with sepsis as BUN levels increased.
Malnutrition is one of the major causes of death in patients with sepsis [27]. Those with sepsis are in a state of severe stress and in a hypercatabolic state, accompanied by organ dysfunction and decreased immunity, making older patients more susceptible to malnutrition. These factors lead to a poor prognosis [28]. The NRS-2002 combines information about a person's nutritional status with clinical disease information. It is a widely used nutritional screening tool for adult inpatients, which can effectively identify critical illnesses with a high nutritional risk. It is not only more sensitive and specific but also more convenient for evaluating malnutrition-related mortality and adverse clinical outcomes of hospitalised patients [29]. Those diagnosed with sepsis are in an imbalanced inflammation-immune state and often show excessive inflammatory reactions. Lipids are peroxidised, proteins are oxidised, and carbohydrates are expended, which increase the risk of malnutrition [30]. Disordered metabolic pathways are critically starved of energy caused by the infection, and this can result in poor clinical outcomes [31]. However, limited data are available regarding NRS-2002 and sepsis [32].
The findings of this study show that the nutritional risk score is an independent risk factor for short-term mortality in older patients with sepsis. With an increase in the nutritional score, the mortality rate of the patients increased significantly in this study. Taking into account the operability of NRS-2002, the screening for malnutrition is now widely recommended for hospitalised patients. Clinicians can quickly evaluate the nutritional risk of these patients, and if they want to optimise sepsis care and prevent iatrogenic malnutrition, targeted nutritional treatment is needed for patients with sepsis. Similarly, some studies have shown that a decreased ALB level is an independent predictor of mortality in patients with sepsis or septic shock, which matches the findings of our study [33]. The mechanism is related to the decrease in colloidal osmotic pressure caused by sepsis, resulting in lower effective circulating volume, decreased ability to bind and transport substances, attenuated antioxidant action, and coagulation dysfunction. Therefore, patients diagnosed with sepsis, especially those at a high risk of undernutrition, should receive individualised nutritional intervention strategies in time, as increasing the level of ALB can improve clinical outcomes.
In conclusion, we analysed the basic data and laboratory indicators of 426 older patients with sepsis in the hospital. This study showed that the indicators ALB, BUN, and Lac levels and NRS-2002 tested for the first time after admission were independent risk factors for poor short-term prognosis (30-day mortality) of older patients with sepsis. We also established a nomogram to predict the prognosis of patients. Practically, there are uncommon models that use nutritional screening scores combined with biomarkers as prognostic factors for older patients with sepsis. The relevant indicators included in this model are convenient and easy to obtain. This model is simple and convenient to evaluate the condition of patients with acute and critical illnesses and provides a reference for clinicians to evaluate the prognosis of patients with sepsis and develop effective interventions.
However, this study also had some shortcomings. First, this study was a single-centre retrospective study, and the number of cases included was relatively small; Only clinical indicators were collected within 24 hours after admission and dynamic change analysis of indicators was not carried out, which could not accurately reflect changes in the patient's condition. Secondly, compared with the prospective clinical design, incomplete case data limited this study, and some meaningful indicators could not be included in the study, such as brain natriuretic peptide levels. Considering the lack of a retrospective analysis to confirm the clinical predictive value of the nomogram constructed in this study, and to provide clinicians with an accurate, reliable, simple, and easy predictive model, the study results needs to be further confirmed by a larger sample-sized study that is a reasonably designed, high-quality prospective study, to obtain a more valuable predictive model for the prognosis of older patients with sepsis.