In our study, we evaluated 119 geriatric patients with PE over 65 years of age. It is known that PESI predicts short-term mortality, which was also validated in our study. At the same time, according to our results, the REMS and HOTEL scores predicted 30 days mortality in geriatric patients with PE as significantly and strongly as PESI. Furthermore, both scores were as powerful as PESI in predicting intensive care requirement in these patients. To the authors’ knowledge, this is the first research to evaluate the prognostic value of HOTEL and REMS in geriatric PE.

In the ED, the workload is intense, and there is a race against time. Therefore, it is crucial that all evaluations provide accurate results and take a short time. SS have thus been developed to support or exclude a diagnosis in patients requiring intensive intervention, such as those with PE [20]. These scores also assist clinicians in determining the severity of emergency cases, refer the patient to the right unit in the post-emergency period, and ensure the efficient use of resources [21].

PESI is a widely used scoring system for PE in all age groups. It consists of 11 independent predictors of disease and clinical state [22]. Although this increases the power of PESI to predict mortality, it also creates a limitation. Three of the parameters included in PESI (history of chronic lung disease, history of cancer, and history of heart failure) are based on anamnesis information. This means that PESI cannot be used in patients whose consciousness level is not sufficient or whose disease history is unknown. This is a major problem in the use of PESI. Aujesky et al. did not make any suggestions to eliminate this limitation when presenting this scoring system [17]. The HOTEL and REMS scores, on the other hand, include some parameters of PESI and are completely based on clinical data. Thus, the consciousness level of the patient or lack of anamnesis is no longer a major problem. There are six parameters in REMS and five parameters in the HOTEL scoring system [23]. Therefore, another advantage of both scores compared to PESI is that they contain fewer parameters. In parallel with this claim, Aujesky et al. also suggested using a simplified version of PESI, which consists of six parameters because it is easier to use. These advantages of HOTEL and REMS, combined with their success in predicting mortality and intensive care requirement, show that they are important potential alternatives to PESI.

In the original study of PESI, Aujesky et al. found the AUC value to be 0.78 (95% CI: 0.77–0.80) in predicting 30 days mortality in patients with PE [17]. In a later meta-analysis, they reported that PESI had an AUC value of 0.7853 (standard deviation: 0.0058) in predicting overall weighted mortality [9]. Despite the small sample size in our study, the AUC values of the three SS were similar in predicting intensive care requirement and mortality. According to our outcomes, the cut-off value of PESI was 111 points, and mortality significantly increased in the patients who scored above this value. Similarly, in the original PESI study including 15,531 patients, this value was reported in high-risk patients (classes III–V) [17].

APACHE II is a well-known scoring system used in the management of critically ill patients. The REMS score, a simplified and easily calculable scoring system, is derived from the more complex APACHE II system [24,25,26,27]. REMS was originally planned to predict the risk of in-hospital mortality [24]. Subsequent studies have shown that REMS strongly predicted poor outcomes in critically ill patients in the ED. In a systematic review, Ghaffarzad et al. reported that REMS successfully predicted 30-day mortality with an AUC value of 0.79 [28]. In another study conducted in the ED with Asian patients, Ha et al. evaluated 1746 non-trauma causes and reported the median REMS score as 6 (IQR: 5–8) and the AUC value as 0.712 for the prediction of 30 days mortality [29]. In a sample including 225 patients with sepsis, Chatchumni et al. observed that REMS had an AUC value of 0.886 and, therefore, suggested that this parameter could be an important predictor of mortality [15]. Our study showed that, REMS had a similar predictive power in predicting mortality. We determined the median REMS as 10, and the AUC value was 0.650. Our slightly lower AUC value can be attributed to our smaller sample size. However, REMS showed better performance in predicting intensive care requirement in our study. Both previous studies and our results suggest that REMS can predict 30 days mortality and intensive care requirement in critically ill patients, such as geriatric patients with PE.

In a previous study, the HOTEL score was reported to be successful in predicting early mortality (15 min–24 h) [18]. Stræde et al. also showed that the HOTEL score had excellent performance in predicting short-term mortality in 1576 medical patients [10]. In another study conducted with 939 geriatric patients in the ED, Dündar et al. found that the HOTEL score strongly predicted intensive care requirement and in-hospital mortality [23]. In the current study, the HOTEL score was as powerful as PESI in predicting short-term mortality and intensive care requirement. Although the HOTEL scoring system was not developed for PE, the parameters it contains are relevant for poor outcomes in this condition. At the same time, the HOTEL score consists of fewer parameters than PESI, which makes it more practical for clinical use in patients with PE. In addition, in the HOTEL score, ECG findings other than a normal sinus rhythm, tachycardia, or bradycardia are considered abnormal. They showed ECG changes in patients with PE in the presence of lobar artery or remote branch and pulmonary trunk involvement [30]. The authors also suggested that ECG would assist clinicians in risk stratification and administration of treatment. Although PESI has 11 parameters, we consider that the similar predictive power of the HOTEL scoring system with only five parameters is due to the inclusion of an ECG evaluation.

Among the patients, 2.5% were discharged, 66.4% were admitted to the inpatient ward, and 31.1% were admitted to the intensive care unit. None of the discharged patients died within 30 days. All mortality was in-hospital mortality. Notably, intensive care unit admission was significantly higher among non-survivors (87.9%) compared to survivors (9.3%), for both inpatient ward and intensive care unit admissions. These results highlight the effectiveness of the scoring system in clinical decision-making, particularly in determining the necessity for intensive care unit admission and risk of death. The significant difference in intensive care unit admissions between survivors and non-survivors underscores the scoring system's ability to accurately predict severe cases requiring intensive care.

This study had certain limitations. The sample size was small. Although this did not prevent significant results, we consider that stronger results can be obtained in larger groups. Second, we were not able to determine whether the abnormal ECG findings existed in the pre-embolism period or developed due to PE. Therefore, all the ECGs with abnormal findings (except sinus tachycardia and sinus bradycardia) were considered abnormal.