https://orcid.org/0009-0006-6701-8354
Figures
Abstract
Background
Several scoring systems have been proposed to predict the risk of death due to severe fever with thrombocytopenia syndrome (STFS), but they have limitations. We developed a new prognostic nomogram for STFS-related death and compared its performance with previous scoring systems and the Acute Physiology and Chronic Health Evaluation score (APACHE II Score).
Methods
A total of 292 STFS patients were retrospectively enrolled between January 2016 and March 2023. Boruta’s algorithm and backward stepwise regression were used to select variables for constructing the nomogram. Time-dependent receiver operating characteristic (ROC) curves and clinical decision curves were generated to compare the strengths of the nomogram with others.
Results
Age, Sequential Organ Failure Assessment Score (SOFA score), state of consciousness, continuous renal replacement therapy (CRRT), and D-dimer were significantly correlated with mortality in both univariate and multivariate analyses (P<0.05). We developed a nomogram using these variables to predict mortality risk, which outperformed the SFTS and APACHE II scores (Training ROC: 0.929 vs. 0.848 vs. 0.792; Validation ROC: 0.938 vs. 0.839 vs. 0.851; P<0.001). In the validation set, the SFTS model achieved an accuracy of 76.14%, a sensitivity of 95.31%, a specificity of 25.00%, a precision of 77.22%, and an F1 score of 85.32%. The nomogram showed a superior performance with an accuracy of 86.36%, a precision of 88.24%, a recall of 93.75%, and an F1 score of 90.91%.
Citation: Yang K, Wang Y, Huang J, Xiao L, Shi D, Cui D, et al. (2024) Establishment and validation of a prognostic nomogram for severe fever with thrombocytopenia syndrome: A retrospective observational study. PLoS ONE 19(10): e0311924. https://doi.org/10.1371/journal.pone.0311924
Editor: Elvan Wiyarta, Universitas Indonesia Fakultas Kedokteran, INDONESIA
Received: May 23, 2024; Accepted: September 26, 2024; Published: October 24, 2024
Copyright: © 2024 Yang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information file.
Funding: This work is funded by Nanjing Second Hospital Reserve Talent Program (0316301), Jiangsu Provincial Health and Health Commission 2021 Medical Research Projects (Jiangsu Health Science and Education (2021) No. 149), 2021 Nanjing Health Science and Technology Development Special Funds Project (YKK21121), Postgraduate Research & Practice Innovation Program of Jiangsu Province (SJCX23_0856) and Nanjing Infectious Disease Clinical Medical Center; Innovation center for infectious disease of Jiangsu Province(NO.CXZX202232). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Severe fever with thrombocytopenia syndrome (SFTS) is a lethal disease caused by severe fever with thrombocytopenia syndrome virus (SFTSV), a member in the Phenuiviridae family and the Bandavirus genus [1]. The virus is transmitted through tick bites or contact with infected body fluids [2–4]. First identified in China in 2009, SFTS has also been reported in Japan, Korea, Vietnam, Myanmar, Taiwan, and Thailand [5–8]. The mortality rate due to SFTS is as high as 30% [2]. The activity of ticks, one important vector of SFTSV, increases with temperature [9]. Thus, the prevalence of SFTS is expected to rise with global warming [10].
SFTS severity is associated with elevated levels of cytokines, such as interleukin-6 (IL-6), IL-8, IL-10, interferon-alpha (IFN-α), and interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α), which are linked to disease progression [11, 12]. In Mainland China, SFTS mainly attacks middle-aged and elderly farmers in rural areas [13]. Currently, however, there are no specific treatments or licensed vaccines for SFTS. In recent years, prognostic indicators based on various risk factors related to SFTS have been explored. However, they have not yielded satisfactory outcomes [14–16]. Our previous all-factor study has shown that the prognosis of SFTS is associated with age, antihypertensive drug application and Acute Physiology and Chronic Health Evaluation Ⅱ (APACHE Ⅱ) score, and a SFTS scoring system has been developed [17]. This system has shown a strong ability to predict the risk of death.
In this prospective observational study, we expanded the sample size, and conducted Boruta algorithm, stepwise regression and multifactorial logistic regression analyses to explore more independent factors. Using these variables, a new nomogram was established, and its accuracy was compared with those of SFTS score and APACHE II score.
Methods
Study population
This study included a total of 292 patients admitted to The Second Hospital of Nanjing over a period from 01/06/2016 to 30/09/2023. Their clinical and laboratory data at admission were retrospectively analyzed. Patients with other virus infections or serious chronic diseases were excluded. SFTS patients were diagnosed based on the presence of acute fever (with a temperature of 38°C or higher) and platelet count <100×10⁹/L), with lab-confirmed SFTS virus (SFTSV) infection by qRT-PCR. We randomized the study population into a derivation set and a validation set (7:3). Their demographics and treatments were balanced between the two sets. The derivation set was divided into survival and death groups. The study was approved by The Second Hospital of Nanjing ethics committees (2023-LS-ky-023). The data were accessed for research purposes at 01/10/2023 to 01/11/2023. Authors did not have access to information that could identify individual participants during or after data collection.
Data collection
In this retrospective study, the clinical and laboratory (blood routine, biochemistry, coagulation) data of the 292 patients diagnosed with SFTS on admission were collected from the Second Hospital of Nanjing Case Data System. The patients were categorized into survival and death groups. Patients’ characteristics, including demographics, medical history, exposure history, comorbidities (e.g., myocardial damage), symptoms, signs, laboratory findings, and treatments (e.g., vasopressors, noninvasive and invasive respiratory support, and renal replacement therapy and primary anti-infective regimen) were recorded. APACHE Ⅱ score and SOFA score were calculated during the first 24 h after admission to the intensive care unit.
Supportive therapies, such as continuous renal replacement therapy, intermittent mandatory ventilation (IMV), vasoactive drugs, ribavirin, hormones, antifungal drugs, antibacterial drugs, were used. Hemofiltration dialysis was set as veno-venous continuous hemodiafiltration (vv-CHDF) at a dose of 25–30 mL/(kg/h). Sodium citrate and calcium gluconate were used for anticoagulation in vitro, with a course of 12 hours. The pH and electrolyte changes were monitored by arterial blood gas every 4 hours. Vasoactive drugs, including norepinephrine, epinephrine, and dobutamine at doses of 0.1–20 mg/(kg/min), were used to maintain average arterial pressure at 65 mmHg. The ventilator was set at modes of assisted/controlled ventilation (A/C), volume control synchronized intermittent mandatory ventilation (VC-SIMV), and pressure control synchronized intermittent mandatory ventilation (PC-SIMV), maintaining peak pressure 40 cm H2O, platform pressure 30 cm H2O, and driving pressure 15 cm H2O and maintaining SpO2 88%.
Pharmacological treatments, such as anti-bacterial and fungal infections, anti-inflammatory and immune-boosting treatments, were applied at the discretion of the doctor depending on the condition of the patient at admission and during hospital stay.
Construction of the nomogram
Firstly, both univariate analysis and the Boruta algorithm were applied to screen significant factors. The factors simultaneously with P<0.05 in the univariate analysis and the non-rejection in the Boruta algorithm were included into the backward stepwise regression analysis to explore the independent factors. Collinearity between continuous variables was tested by the variance inflation factor (VIF), and an arithmetic square root of VIF ≤2 was considered as non-collinearity. MeanDecreaseGini and MeanDecreaseAccuracy analyses were used to visualize the importance of each factor. Finally, the results of the backward stepwise regression analysis were subjected to the binary logistic regression analysis to explore the independent factors affecting the prognosis of SFTS. Based on them, a nomogram was obtained in the training set according to Occam’s Law of Razor. The best model should be one that could achieve the best performance with the least variables [18].
Validation of the nomogram
The performances of the nomogram, SFTS score and APACHE II score were compared in terms of area under curve (AUC), integrated discrimination improvement (IDI), decision curve analysis (DCA) in both the training and validation sets. In the validation set, the total accuracy (TA) and consistency (Kappa coefficient) of the nomogram and SFTS score were compared.
In both training and validation sets, the performance of the nomogram was evaluated by an area under the curve of the receiver operating characteristic (AUROC) and by calibration with bootstrap method with 1000 resampling. Hosmer and Lemeshow Goodness of Fit (GOF) tests were applied to model calibration evaluation. DCA analysis was performed to evaluate the net benefit of medical intervention conforming nomogram.
Statistical analysis
Statistical analyses were performed using the R software (version4.3.1 R Foundation for Statistical Computing, Vienna, Austria). In the univariate analysis, categorical data were analyzed by x2 test. Continuous data in a normal distribution were analyzed by independent t test and presented as . Continuous data in a non-normal distribution were analyzed by the Mann Whitney test and their medians were compared (P25, P75).
Missing values were addressed with multiple imputation in the process of logistic regression and model construction with five interpolations. The imputation technique involved creating multiple copies of the data and replacing missing values with imputed values through a suitable random sample from their predicted distribution. A two-tailed p value <0.05 was considered statistically significant. Candidate variables for multifactor analysis were screened using Brouta’s algorithm in conjunction with univariate analysis, where green was significant, yellow was uncertain, and red was rejected in Brouta’s algorithm visualization graph. Variables associated with death in the univariate analyses (p < 0.05) were included into the multivariate logistic regression analysis, and their estimated odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated, with a significance level of p < 0.05 for independent risk factors. We performed backward stepwise regression analysis on the dataset using the R package "stats". All analyses were reported according to the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines [19].
Results
Patient description
Table 1 shows the baseline characteristics of patients at admission in the training and validation sets. The total patients were randomly grouped in a 7:3 ratio into the training and validation sets. Baseline characteristics data, except for Mean arterial pressure (MAP), were well balanced between the two cohorts. Between 2016 and 2023, a total of 302 patients had positive qRT-PCR results for SFTS virus. Cases with missing data or who died within 24 hours after admission were excluded. Then, 292 patients were finally involved (Fig 1), with a median age of 68 years (range 58–72 years). The demographic and pathological features of the patients are summarized in Table 1.
Table 2 shows that 48 of the 204 patients in the Derivation set died, with a mortality rate of 23.5%. The gender difference between the two groups was not significant. The median age of patients was 70.5 (64.75, 74.25) in the death group and 67 (56, 71.25) in the survival group. The demographic and pathological features in the training set are summarized in Table 2
Development of a prediction nomogram in the training set
The Boruta algorithm was applied to select all the features associated with prognosis, and the Maximum Z Score among Shadowed Attributes (MZSA) was found to distinguish significant from non-significant features. A total of 12 variables (consciousness, APACHE Ⅱ Score, SOFA Score, heart rate (HR), vasopressors, medicine ventilator (MV), CRRT, blood urea nitrogen (BUN), aspartate transferase (AST), lactate dehydrogenase (LDH), creatinine (Cr), D-Dimer) were identified (green) as important features, three variables (antifungal, viral load, age) as pending, and the rest as rejected (red) (Fig 2A). The stabilization of variation of Z-scores during the Boruta run was shown (Fig 2B). Of the above 15 variables, 6 significant were filtered out in a backward stepwise regression analysis (age, consciousness, SOFA Score, vasopressors, CRRT, D-Dimer) and further subjected to a binary logistic regression analysis, further showing that age, consciousness, SOFA Score, vasopressors, CRRT, D-Dimer as independent risk factors for poor prognosis in SFTS patients (Table 3). The VIF was calculated, and all variables in Fig 3A had an arithmetic square root of VIF ≤2. MeanDecreaseGini and MeanDecreaseAccuracy analyses indicated that the above five independent risks (Fig 3B) could be used as key variables for predicting the prognosis of patients with SFTS, and funneled into the nomogram scoring system (Fig 3C).
(A) Boruta variable selection graph; (B) Boruta iterative trajectory diagram.
(A) VIF values of less than 2 for each independent influencing factor; (B) MeanDecreaseGini and MeanDecreaseAccuracy analyses; (C) Nomogram scoring system. CRRT = Continuous Renal Replacement Therapy; SOFA score = Sequential Organ Failure Assessment Score; VIF = Variance Inflation Factor.
To apply the nomogram for prognosis prediction, patients’ data, including age, consciousness, SOFA score, CRRT, and D-Dimer levels, were first collected. Following this, the total score was calculated by summing the points assigned to each of these factors according to the model. Finally, the total score was used to assess the patient’s mortality risk, allowing for a nuanced evaluation of their prognosis.
Comparison between three scoring systems
We compared the nomogram with SFTS score and APACHE Ⅱ score in predicting the mortality in SFTS patients. As shown in Table 4, the AUROC of the nomogram was significantly higher than those of SFTS Score and APACHE Ⅱ Score in both the training and validation sets, indicating that the nomogram had a stronger discriminative ability (Fig 4).
(A) Training set; (B) Validation set.
Calibration curves in both the training and validation sets revealed a good agreement between the actual and predicted probabilities (Fig 5). The values of Hosmer and Lemeshow Goodness of Fit (GOF) tests were 0.612 and 0.87 in the training and validation sets, respectively. The IDI of the nomogram scoring system was significantly higher than those of SFTS Score and APACHE Ⅱ Score in both sets, revealing that the nomogram could increase the predictive accuracy of SFTS Score and APACHE Ⅱ (Table 4). In the validation group, the accuracies of the nomogram and SFTS score were 86.4% (95% CI: 0.774, 0.928; P = 0.002) and 76.1% (95% CI: 0.659, 0.846; P = 0.278), respectively (Fig 6). The sensitivity and specificity of the nomogram were 93.8% and 66.7%, and those of SFTS score were 95.3% and 25.0%, respectively.
(A) Training set; (B) Validation set.
(A) Nomogram; (B) SFTS score.
The DCA for the nomogram was depicted and compared with those of SFTS score and APACHE Ⅱ score. In the training set, medical intervention guided by the SFTS score could add more net benefit than the nomogram and APACHE Ⅱ score, when the threshold probability (PT) was between 0.1 and 0.73. Nevertheless, the nomogram could add more net benefit than SFTS score and APACHE Ⅱ Score when the PT was between 0.73 and 0.1 (Fig 7A). In the validation set, treatment directed by SFTS score could gain more net benefit than the nomogram and APACHE Ⅱ Score, when the PT was between 0.1 and 0.6 (Fig 7B), but the nomogram could add more net benefit than the SFTS and APACHE Ⅱ Score when the PT was between 0.61 and 0.71(Fig 7B).
(A) Training set; (B) Validation set.
Discussion
By analyzing the clinical data of patients SFTS admitted at Nanjing Second Hospital (2016–2023), we conducted logistic regression analyses to recognize the risk factors related to mortality in STFS patients, including age, consciousness, SOFA score, CRRT and D-Dimer. Based on them, the nomogram achieved a satisfactory performance in predicting the mortality.
Nomograms allow researchers to visualize the prognostic outcomes of patients directly and quickly. Efficient nomograms need to be established on stable and significant variables, which are traditionally filtered out by a combination of single-factor and multi-factor analyses. However, these analyses have certain drawbacks, such as ignorance on interaction between factors, lack of a holistic view, existence of under-interpretation [15, 16]. Boruta’s algorithm has high accuracy and robustness to identify important features by using shadow features and an iterative process [20]. Compared to other methods, Boruta’s algorithm is superior in processing high-dimensional and nonlinear data, and can be integrated with other models [21]. We utilized the Boruta’s algorithm (Fig 2) in conjunction with backward stepwise regression analysis for feature selection on high-dimensional data, and screened out age, consciousness, SOFA Score, CRRT and D-Dimer (all with VIF<2) as independent factors (Fig 3A).
The present study showed that the mortality rate was as high as 24.7% in the total population, and the age of the patients in the death group was higher than that in the survivor group (Table 2). Age was an independent risk factor for prognosis (OR = 1.11, 95%CI: 1.04,1.18) (P = 0.002) (Table 3). There was no statistically significant difference between the two groups in terms of gender, which is consistent with the findings of Gong L, Yokomizo K and Chen Q, et al. [22–24]. An elder age indicated a higher mortality, which may be due to the fact that most elderly patients have a combination of underlying diseases, an increased risk of severe infections, a low organ compensatory capacity, and an impaired immunity [25, 26]. The SOFA score can be used to measure sepsis-related organ dysfunction in sepsis-related studies [27, 28], and perform well in assessing the severity of infectious diseases [29–31]. SFTV infection is mainly manifested as high fever, thrombocytopenia, leukocytopenia, gastrointestinal symptoms, liver and kidney function damage, and even MODS [32]. Karakike E et al. have noted that SOFA score has a high clinical value in predicting 28-day prognosis of infectious diseases [33]. Similarly, our study showed that the risk of death in patients with a high SOFA score was 1.58 times higher than that in patients with a low score, and our multifactorial analysis showed that SOFA score was an independent risk factor for the prognosis in patients with SFTS (OR = 1.58, 95% CI:1.24,2.01) (P<0.001) (Table 3). MeanDecreaseGini and MeanDecreaseAccuracy analyses indicated that SOFA score (Fig 3B) was the most important key variable in predicting the prognosis of patients with SFTS. Therefore, dynamic monitoring of SOFA score may be of great clinical value in assessing disease progression.
Consciousness is a component of the SOFA score. Studies have shown high morbidity and mortality in SFTS-related encephalitis or encephalopathy [34]. CNS manifestations (adjusted odds ratio [OR] 30·26) are important risk factors for the fatal outcomes of SFTS patients [35, 36]. Similarly, the present study showed that patients with altered consciousness had a 4.71 times higher risk of death than those without, indicating it as an independent prognostic factor for SFTS (OR = 4.71, 95% CI:1.42,14.14) (P = 0.011) (Table 3). The pathogenic mechanism of encephalitis in patients with SFTS is unclear, and studies have shown that the virus is able to cross the blood-brain barrier into the skull to replicate, thus directly damaging neurons in brain tissue [37, 38]. D-dimer level has been the focus of studies about Coronavirus Disease 2019 (COVID-19) during the pandemic. Hypercoagulability may be a key mechanism for acute organ injury and death in COVID-19 patients. A higher D-dimer level has been found independently associated with a higher risk of death in a large multicenter cohort study of critically ill patients with COVID-2019 [39]. Meanwhile, coagulation abnormalities are also common in patients with STFS. Studies have shown that coagulation parameters are more sensitive than platelet count in early prediction of SFTS, with D-dimer levels much higher in the death group than in the survival group [40]. Similarly, the present study showed that the platelet count was lower in the death group than in the survival group at admission (P<0.001), but not an independent predictive factor for mortality; whereas, the D-dimer level (4.24 [1.94,13.92]) was higher than that in the survival group (1.77[0.82,4.49]) (P< 0.001), and was identified as an independent risk factor (OR = 1.05, 95%CI:1.01,1.09) (P = 0.006) in the multifactorial analysis (Table 3). Renal damage is one of the MOFs. Studies have shown that AKI is associated with poor outcomes in patients with SFTS, especially a high mortality in patients with AKI stage 2 or 3 [41]. In this study, we showed that the mortality rate was higher in patients who required CRRT at admission than those who did not, and the multifactorial analysis showed that the mortality was statistically 4.58 times higher in patients who required CRRT treatment than those who did not (P = 0.042) (Tables 2, 3).
Based on the five independent risk factors mentioned above, we constructed a prediction nomogram (Fig 3C), which achieved an accuracy of 86.36% in the validation set (Fig 6A). Currently, several scoring systems, such as APACHE score, SOFA score, and Simplified Acute Physiology Score II, have been proposed for prediction in ICU patients. APACHE II is the most widely used. Here, we further showed that the AUROC of the nomogram was 0.929 in the training set and 0.938 in the validation set, both higher than those of APACHE II score and the STFS score in our previous study (P<0.05), suggestive of its excellent discriminative ability (Fig 4A and 4B). In terms of IDI, the nomogram outperformed APACHE II score and STFS score (Table 4). The APACHE II score performed the worst, indicating that it may be not suitable to predict multiple organ failure [42].
This study has several advantages compared with the previous. This study validated that SFTS score system also had a good predictive ability and a high accuracy (training set: AUC = 0.848, validation set: AUC = 0.851) (Fig 4B), and surprisingly, the net clinical benefit of SFTS score was within a larger threshold range than the nomogram and APACHE II score (Fig 7A and 7B), and remained consistent across the training and validation sets. The net clinical benefit of the nomogram was higher than that of SFTS score within a certain threshold range. Therefore, we propose that both scoring systems may be complementary in clinical prediction of disease outcomes.
This study is limited by a single-center sample, a small validation set, and an insufficient generality. Due to the small sample size, the model may only be able to learn specific features in the training set and miss other meaningful features. Second, the subjects of the retrospective study may not be fully representative, and unidentified confounding factors may have biased the results.
In conclusion, the nomogram based on five factors (age, consciousness, SOFA score, CRRT and D-Dimer) showed an excellent performance in predicting the prognosis of STFS patients. The STFS score system, developed by us in the previous studies, has also a good predictive ability. Both are expected to be used in more clinical practices.
References
- 1. Seo JW, Kim D, Yun N, Kim DM. Clinical Update of Severe Fever with Thrombocytopenia Syndrome. Viruses. 2021;13(7). Epub 2021/07/03. pmid:34201811; PubMed Central PMCID: PMC8310018.
- 2. Yu XJ, Liang MF, Zhang SY, Liu Y, Li JD, Sun YL, et al. Fever with thrombocytopenia associated with a novel bunyavirus in China. N Engl J Med. 2011;364(16):1523–32. Epub 2011/03/18. pmid:21410387; PubMed Central PMCID: PMC3113718.
- 3. Tian H, Yu P, Chowell G, Li S, Wei J, Tian H, et al. Severe Fever with Thrombocytopenia Syndrome Virus in Humans, Domesticated Animals, Ticks, and Mosquitoes, Shaanxi Province, China. Am J Trop Med Hyg. 2017;96(6):1346–9. Epub 2017/07/19. pmid:28719252; PubMed Central PMCID: PMC5462569.
- 4. Wu YX, Yang X, Leng Y, Li JC, Yuan L, Wang Z, et al. Human-to-human transmission of severe fever with thrombocytopenia syndrome virus through potential ocular exposure to infectious blood. Int J Infect Dis. 2022;123:80–3. Epub 2022/08/21. pmid:35987469.
- 5. Takahashi T, Maeda K, Suzuki T, Ishido A, Shigeoka T, Tominaga T, et al. The first identification and retrospective study of Severe Fever with Thrombocytopenia Syndrome in Japan. J Infect Dis. 2014;209(6):816–27. Epub 2013/11/16. pmid:24231186; PubMed Central PMCID: PMC7107388.
- 6. Lin TL, Ou SC, Maeda K, Shimoda H, Chan JP, Tu WC, et al. The first discovery of severe fever with thrombocytopenia syndrome virus in Taiwan. Emerging microbes & infections. 2020;9(1):148–51. Epub 2020/01/11. pmid:31918622; PubMed Central PMCID: PMC6968498.
- 7. Tran XC, Yun Y, Van An L, Kim SH, Thao NTP, Man PKC, et al. Endemic Severe Fever with Thrombocytopenia Syndrome, Vietnam. Emerg Infect Dis. 2019;25(5):1029–31. Epub 2019/04/20. pmid:31002059; PubMed Central PMCID: PMC6478219.
- 8. Li JC, Wang YN, Zhao J, Li H, Liu W. [A review on the epidemiology of severe fever with thrombocytopenia syndrome]. Zhonghua liu xing bing xue za zhi = Zhonghua liuxingbingxue zazhi. 2021;42(12):2226–33. Epub 2021/12/27. pmid:34954991.
- 9. Kim M, Heo ST, Oh H, Oh S, Lee KH, Yoo JR. Prognostic Factors of Severe Fever with Thrombocytopenia Syndrome in South Korea. Viruses. 2020;13(1). Epub 2020/12/31. pmid:33374620; PubMed Central PMCID: PMC7822455.
- 10. Noh JY, Song JY, Bae JY, Park MS, Yoon JG, Cheong HJ, et al. Seroepidemiologic survey of emerging vector-borne infections in South Korean forest/field workers. PLoS Negl Trop Dis. 2021;15(8):e0009687. Epub 2021/08/19. pmid:34407077; PubMed Central PMCID: PMCPMC8405005.
- 11. Song L, Zou W, Wang G, Qiu L, Sai L. Cytokines and lymphocyte subsets are associated with disease severity of severe fever with thrombocytopenia syndrome. Virol J. 2024;21(1):126. Epub 2024/06/04. pmid:38831352; PubMed Central PMCID: PMC11149350.
- 12. Hu LF, Wu T, Wang B, Wei YY, Kong QX, Ye Y, et al. The Regulation of Seventeen Inflammatory Mediators are Associated with Patient Outcomes in Severe Fever with Thrombocytopenia Syndrome. Sci Rep. 2018;8(1):159. Epub 2018/01/11. pmid:29317732; PubMed Central PMCID: PMC5760584.
- 13. Huang X, Li J, Li A, Wang S, Li D. Epidemiological Characteristics of Severe Fever with Thrombocytopenia Syndrome from 2010 to 2019 in Mainland China. Int J Environ Res Public Health. 2021;18(6). Epub 2021/04/04. pmid:33802869; PubMed Central PMCID: PMC8002760.
- 14. Xiong S, Zhang W, Li M, Xiong Y, Li M, Wang H, et al. A simple and practical score model for predicting the mortality of severe fever with thrombocytopenia syndrome patients. Medicine (Baltimore). 2016;95(52):e5708. Epub 2016/12/30. pmid:28033271; PubMed Central PMCID: PMC5207567.
- 15. Wang L, Wan G, Shen Y, Zhao Z, Lin L, Zhang W, et al. A nomogram to predict mortality in patients with severe fever with thrombocytopenia syndrome at the early stage-A multicenter study in China. PLoS Negl Trop Dis. 2019;13(11):e0007829. Epub 2019/11/26. pmid:31765414; PubMed Central PMCID: PMC6934327.
- 16. Jia B, Yan X, Chen Y, Wang G, Liu Y, Xu B, et al. A scoring model for predicting prognosis of patients with severe fever with thrombocytopenia syndrome. PLoS Negl Trop Dis. 2017;11(9):e0005909. Epub 2017/09/22. pmid:28934195; PubMed Central PMCID: PMC5626493.
- 17. Yang K, Chen J, Chen Z, Zheng Y. Risk Factors for Death in Patients with Severe Fever with Thrombocytopenia Syndrome. Am J Trop Med Hyg. 2023;109(1):94–100. Epub 2023/05/31. pmid:37253446; PubMed Central PMCID: PMC10324000.
- 18. Van Den Berg HA. Occam’s razor: from Ockham’s via moderna to modern data science. Sci Prog. 2018;101(3):261–72. Epub 2018/07/22. pmid:30025552; PubMed Central PMCID: PMC10365162.
- 19. Collins GS, Dhiman P, Andaur Navarro CL, Ma J, Hooft L, Reitsma JB, et al. Protocol for development of a reporting guideline (TRIPOD-AI) and risk of bias tool (PROBAST-AI) for diagnostic and prognostic prediction model studies based on artificial intelligence. BMJ open. 2021;11(7):e048008. Epub 2021/07/11. pmid:34244270; PubMed Central PMCID: PMC8273461.
- 20. Kursa MB, Rudnicki WR. Feature Selection with the Boruta Package. Journal of Statistical Software. 2010;36(11):1–13.
- 21. Costa OYA, de Hollander M, Pijl A, Liu B, Kuramae EE. Cultivation-independent and cultivation-dependent metagenomes reveal genetic and enzymatic potential of microbial community involved in the degradation of a complex microbial polymer. Microbiome. 2020;8(1):76. Epub 2020/06/03. pmid:32482164; PubMed Central PMCID: PMC7265232.
- 22. Gong L, Zhang L, Wu J, Lu S, Lyu Y, Zhu M, et al. Clinical Progress and Risk Factors for Death from Severe Fever with Thrombocytopenia Syndrome: A Multihospital Retrospective Investigation in Anhui, China. Am J Trop Med Hyg. 2021;104(4):1425–31. Epub 2021/02/17. pmid:33591933; PubMed Central PMCID: PMC8045620.
- 23. Yokomizo K, Tomozane M, Sano C, Ohta R. Clinical Presentation and Mortality of Severe Fever with Thrombocytopenia Syndrome in Japan: A Systematic Review of Case Reports. Int J Environ Res Public Health. 2022;19(4). Epub 2022/02/26. pmid:35206459; PubMed Central PMCID: PMC8871924.
- 24. Chen Q, Yang D, Zhang Y, Zhu M, Chen N, Yushan Z. Transmission and mortality risk assessment of severe fever with thrombocytopenia syndrome in China: results from 11-years’ study. Infectious diseases of poverty. 2022;11(1):93. Epub 2022/09/05. pmid:36058928; PubMed Central PMCID: PMC9440863.
- 25. Liang S, Xie W, Li Z, Zhang N, Wang X, Qin Y, et al. Analysis of fatal cases of severe fever with thrombocytopenia syndrome in Jiangsu province, China, between 2011 and 2022: A retrospective study. Frontiers in public health. 2023;11:1076226. Epub 2023/04/11. pmid:37033043; PubMed Central PMCID: PMC10076888.
- 26. Gai ZT, Zhang Y, Liang MF, Jin C, Zhang S, Zhu CB, et al. Clinical progress and risk factors for death in severe fever with thrombocytopenia syndrome patients. J Infect Dis. 2012;206(7):1095–102. Epub 2012/08/02. pmid:22850122.
- 27. Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A, et al. Assessment of Clinical Criteria for Sepsis: For the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). JAMA. 2016;315(8):762–74. Epub 2016/02/24. pmid:26903335; PubMed Central PMCID: PMC5433435.
- 28. Kilinc Toker A, Kose S, Turken M. Comparison of SOFA Score, SIRS, qSOFA, and qSOFA + L Criteria in the Diagnosis and Prognosis of Sepsis. The Eurasian journal of medicine. 2021;53(1):40–7. Epub 2021/03/16. pmid:33716529; PubMed Central PMCID: PMC7929579.
- 29. McBride A, Vuong NL, Van Hao N, Huy NQ, Chanh HQ, Chau NTX, et al. A modified Sequential Organ Failure Assessment score for dengue: development, evaluation and proposal for use in clinical trials. BMC Infect Dis. 2022;22(1):722. Epub 2022/09/04. pmid:36057771; PubMed Central PMCID: PMC9441074.
- 30. Raschke RA, Agarwal S, Rangan P, Heise CW, Curry SC. Discriminant Accuracy of the SOFA Score for Determining the Probable Mortality of Patients With COVID-19 Pneumonia Requiring Mechanical Ventilation. JAMA. 2021;325(14):1469–70. Epub 2021/02/18. pmid:33595630; PubMed Central PMCID: PMC7890534.
- 31. Teparrukkul P, Hantrakun V, Imwong M, Teerawattanasook N, Wongsuvan G, Day NP, et al. Utility of qSOFA and modified SOFA in severe malaria presenting as sepsis. PLoS One. 2019;14(10):e0223457. Epub 2019/10/10. pmid:31596907; PubMed Central PMCID: PMC6785116.
- 32. Luo N, Li M, Xu M, Shi C, Shi X, Ni R, et al. Research Progress of Fever with Thrombocytopenia Syndrome. Intensive care research. 2023:1–10. Epub 2023/06/26. pmid:37360310; PubMed Central PMCID: PMC10033304.
- 33. Karakike E, Kyriazopoulou E, Tsangaris I, Routsi C, Vincent JL, Giamarellos-Bourboulis EJ. The early change of SOFA score as a prognostic marker of 28-day sepsis mortality: analysis through a derivation and a validation cohort. Crit Care. 2019;23(1):387. Epub 2019/12/01. pmid:31783881; PubMed Central PMCID: PMC6884794.
- 34. Xu Y, Shao M, Liu N, Dong D, Tang J, Gu Q. Clinical feature of severe fever with thrombocytopenia syndrome (SFTS)-associated encephalitis/encephalopathy: a retrospective study. BMC Infect Dis. 2021;21(1):904. Epub 2021/09/05. pmid:34479504; PubMed Central PMCID: PMC8418043.
- 35. Li H, Lu QB, Xing B, Zhang SF, Liu K, Du J, et al. Epidemiological and clinical features of laboratory-diagnosed severe fever with thrombocytopenia syndrome in China, 2011–17: a prospective observational study. Lancet Infect Dis. 2018;18(10):1127–37. Epub 2018/07/29. pmid:30054190.
- 36. Wang X, Ren X, Ge Z, Cui S, Wang L, Chen Z, et al. Clinical manifestations of death with severe fever and thrombocytopenia syndrome: A meta-analysis and systematic review. J Med Virol. 2021;93(6):3960–8. Epub 2020/09/16. pmid:32930400.
- 37. Cui N, Liu R, Lu QB, Wang LY, Qin SL, Yang ZD, et al. Severe fever with thrombocytopenia syndrome bunyavirus-related human encephalitis. J Infect. 2015;70(1):52–9. Epub 2014/08/20. pmid:25135231.
- 38. Chen R, Li Q, Chen H, Yang H, Wei X, Chen M, et al. Severe fever with thrombocytopenia syndrome virus replicates in brain tissues and damages neurons in newborn mice. BMC Microbiol. 2022;22(1):204. Epub 2022/08/21. pmid:35987890; PubMed Central PMCID: PMC9392058.
- 39. Short SAP, Gupta S, Brenner SK, Hayek SS, Srivastava A, Shaefi S, et al. d-dimer and Death in Critically Ill Patients With Coronavirus Disease 2019. Crit Care Med. 2021;49(5):e500–e11. Epub 2021/02/17. pmid:33591017; PubMed Central PMCID: PMC8275993.
- 40. Xu L, Wang H, Liang BY, Wang T, Zheng X, Peng C. [Clinical features and significance of coagulation dysfunction in severe fever with thrombocytopenia syndrome]. Zhonghua Nei Ke Za Zhi. 2022;61(7):793–6. Epub 2022/06/29. pmid:35764563.
- 41. Zhang Z, Hu X, Jiang Q, Hu W, Li A, Deng L, et al. Clinical characteristics and outcomes of acute kidney injury in patients with severe fever with thrombocytopenia syndrome. Front Microbiol. 2023;14:1236091. Epub 2023/10/02. pmid:37779695; PubMed Central PMCID: PMC10533938.
- 42. Bahtouee M, Eghbali SS, Maleki N, Rastgou V, Motamed N. Acute Physiology and Chronic Health Evaluation II score for the assessment of mortality prediction in the intensive care unit: a single-centre study from Iran. Nurs Crit Care. 2019;24(6):375–80. Epub 2019/03/30. pmid:30924584.