Prognostic factors for the short-term mortality of patients with rheumatoid arthritis admitted to intensive care units

Explainable machine learning model for prediction of 28-day all-cause mortality in immunocompromised patients in the intensive care unit: a retrospective cohort study based on MIMIC-IV database

OBJECTIVES

This study aimed to develop and validate an explainable machine learning (ML) model to predict 28-day all-cause mortality in immunocompromised patients admitted to the intensive care unit (ICU). Accurate and interpretable mortality prediction is crucial for clinical decision-making and optimal allocation of critical care resources for this vulnerable patient population.

METHODS

We utilized retrospective clinical data from the MIMIC-IV (version 2.2) database, encompassing ICU admissions at Beth Israel Deaconess Medical Center from 2008 to 2019. Eligible immunocompromised patients, including those with primary immunodeficiencies and chronic acquired conditions, such as hematological malignancies, solid tumors, and organ transplantation, were selected. Data were randomly split into training (80%) and testing (20%) cohorts. Ten ML models (logistic regression, XGBoost, LightGBM, AdaBoost, Random Forest, Gradient Boosting, Gaussian Naive Bayes, Complement Naive Bayes, Multilayer Perceptron, and Support Vector Machine) were developed and evaluated using area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), sensitivity, specificity, accuracy, and F1 score. Model explainability was achieved through SHapley Additive exPlanations (SHAP), and decision curve analysis (DCA) assessed clinical utility. In addition, Cox proportional hazards regression was conducted to evaluate the impact of predictive factors on time-to-event outcomes.

RESULTS

Among the evaluated models, the Support Vector Machine (SVM) demonstrated the highest AUROC of 0.863 (95% CI 0.834-0.890) and a notable AUPRC of 0.678 (95% CI 0.624-0.736). Key predictive factors consistently identified across multiple ML models included 24-h urine output, blood urea nitrogen (BUN) levels, presence of metastatic solid tumors, Charlson Comorbidity Index (CCI), and international normalized ratio (INR). SHAP analyses provided detailed insights into how these features influenced model predictions.

CONCLUSIONS

The explainable ML models based on various artificial intelligence methods demonstrated promising clinical applicability in predicting 28-day mortality risk among immunocompromised ICU patients. Factors such as urine output, BUN, metastatic solid tumors, CCI, and INR significantly contributed to prediction outcomes and may serve as important predictors in clinical practice.

Association of calcaneal pitch angle with recurrence of postoperative hallux valgus in patients with rheumatoid arthritis

Hallux valgus (HV) and flatfoot deformities are frequently seen in patients with rheumatoid arthritis (RA). This study aimed to determine whether flatfoot deformity contributes to the recurrence of HV in RA patients. This study examined 62 feet from 45 RA patients who were diagnosed with HV and underwent the first metatarsal joint-preserving surgery between November 2010 and October 2021. Recorded data included age at surgery, sex, disease duration, body mass index [BMI], RA disease duration, medical treatment of RA, Larsen grade, blood test, pre/postoperative Japanese Society for Surgery of the Foot, HV angle, M1M2 angle, M1M5 angle, calcal pitch angle, and Meary's angle. HV recurrence on radiography was defined as an HV angle exceeding 20°. HV recurrence was observed in 17 feet. Significant differences were observed due to risk factors such as BMI, disease duration, Larsen grade 4-5, and preoperative calcaneal pitch angle. Multivariate logistic regression analysis identified that lower BMI, a higher M1M2 angle, and a lower calcaneal pitch angle are preoperative risk factors for the recurrence of postoperative HV in RA patients. LEVEL OF EVIDENCE: 3.

Decreased in-hospital mortality in shock patients with rheumatoid arthritis compared to those without: a retrospective analysis of a critical care database

Inflammatory response is a natural protective response in infection or injury. The immune imbalance of rheumatoid arthritis (RA) patients and the use of anti-rheumatic immune drugs may affect the inflammatory response process of shock. The evidence of the impact of RA on short-time mortality in shock patients is limited. This study was a retrospective observational study from the Medical Information Mart for Intensive Care IV database (MIMIC-IV). Patients diagnosed with shock were selected from the database. The primary outcome was in-hospital mortality. Binary logistic regression analysis was used to explore the association between RA and in-hospital mortality. Subgroup analysis of the association between in-hospital mortality and RA was performed in septic shock, cardiogenic shock, and other types of shock. A total of 8233 ICU patients with shock were identified. After adjusting for confounding variables, RA was associated with a lower risk of in-hospital mortality (OR, 0.58; 95% CI, 0.38-0.91). The results of the subgroup analysis showed that RA was associated with a decreased risk of in-hospital death in the septic shock group but not in cardiogenic shock or other types of shock. RA was associated with a lower risk of in-hospital death in shock patients. This association was only observed in septic shock but not in cardiogenic shock and other types of shock. Key Points • Rheumatoid arthritis (RA) is associated with a lower risk of in-hospital mortality in ICU patients with shock, particularly in those with septic shock. • This study utilized propensity score matching to accurately compare outcomes between RA and non-RA shock patients, ensuring balanced baseline characteristics. • Findings suggest that the immune modulation inherent to RA or the impact of anti-rheumatic drugs might influence the survival outcomes in shock scenarios. • The protective effect of RA on in-hospital mortality was not observed in patients with cardiogenic shock or other types of shock, indicating a specific interaction in septic shock contexts.

First Metatarsophalangeal Joint-Preserving Surgery Is Effective for Forefoot Deformity With Moderate to Severe Joint Destruction in Rheumatoid Arthritis

Background

The study compares the clinical outcomes of first metatarsophalangeal (MTP) joint fusion vs joint-preserving surgery in rheumatoid arthritis (RA) patients with severe forefoot deformities.

Methods

This single-center retrospective study at Kyushu University Hospital reviewed RA patients who underwent either first MTP joint arthrodesis or joint-preserving surgery for hallux valgus (HV) deformity between January 2008 and December 2022. A total of 103 feet (73 cases) were analyzed, with 75 feet (58 cases) showing radiographic bone destruction of Larsen grade 3 or higher. One foot underwent resection arthroplasty, so ultimately 74 feet in 57 cases were evaluated. Surgical procedures included joint-preserving biplane osteotomy or arthrodesis with crossed screws. Clinical outcomes were measured using the Japanese Society for Surgery of the Foot (JSSF) Hallux scale, whereas radiographic assessments included HV angle (HVA). Propensity score matching was used to minimize bias when comparing postoperative outcomes between the arthrodesis and joint-preserving surgery groups.

Results

This study analyzed 74 feet undergoing either arthrodesis (27 feet) or joint-preserving surgery (47 feet) for HV. Patients in the 2 groups showed similar demographic and clinical characteristics except with respect to length of follow-up, which was greater in the arthrodesis group (5.1 ± 2.6 years vs 2.4 ± 2.0 years, P < .01) than the joint-preserving group. In the arthrodesis group, all patients underwent resection arthroplasty on the second to fifth toes. The joint-preserving group included first MTP joint surgery alone (n = 5) and first MTP joint and lesser MTP joint surgeries (resection arthroplasty, n = 29; joint-preserving surgery, n = 13). Functional scores significantly improved in both groups, with first metatarsophalangeal joint-preserving surgery yielding better postoperative outcomes. In cases of deformity recurrence, the recurrent cases exhibited greater immediate postsurgical HVA, but other foot function outcomes remained similar at the end of follow-up.

Conclusion

Joint-preserving surgery for advanced rheumatoid forefoot deformity showed better functional improvement than arthrodesis using the propensity score matching and comparable clinical outcomes, highlighting it as a potential treatment option for severe joint destruction.

Level of Evidence

Level Ⅳ, retrospective study.

A Review of the Current Status of Anesthetic Management of Patients with Rheumatoid Arthritis

Rheumatoid arthritis (RA) is a chronic connective tissue disease of immunological etiology. In the course of the disease, symptoms of the musculoskeletal system predominate, but other systems can also be affected. The disease may require long-term treatment, and patients often require surgery on damaged joints. Complications of the disease and drug interactions may contribute to difficulties in perioperative care; therefore, knowledge is required to provide appropriate care. When anesthetizing a patient with RA, we should pay special attention to preoperative evaluation, taking a medical history, risk of difficult intubation or cardiac incidents, respiratory insufficiency, and frequent pulmonary infections. It is important to be aware of perioperative glucocorticoids supplementation, especially in patients with suspected adrenal insufficiency. Postoperative management, such as pain management, early rehabilitation, and restart of pharmacotherapy play, an important role in the patient's recovery. Special attention should be paid to perioperative management in pregnant women, as the disease is a significant risk factor for complications, and some anesthetic procedures can be noxious to the fetus. Due to the nature of the disease, it can be challenging for the anesthesiologist to provide good and appropriate pain medications, symptom management, and other necessary techniques that are done to anesthetize the patient properly. This work is based on the available literature and the authors' experience. This article aims to review the current status of anesthetic management of patients with rheumatoid arthritis.

Mortality Trends Among Patients with Rheumatoid Arthritis in Western Australia

INTRODUCTION

With scarce comparative data on mortality in Australian patients with rheumatoid arthritis (RA), we investigated temporal changes in standardized mortality rates for patients with RA using longitudinal linked population-wide health data in Western Australia (WA) over the period 1980 to 2015.

METHODS

The study included 17,125 patients with a first-time hospital contact for RA (ICD-10-AM M05.00-M06.99 and ICD-9-AM 714.00-714.99) in the study period. Standardized mortality rate ratios (SMRRs) for the RA cohort versus the WA general population was estimated using direct age standardization. We analyzed temporal trends over with dates and causes provided by the WA Death Registry.

RESULTS

During 356,069 patient-years of follow-up, a total of 8955 (52%) deaths occurred in the RA cohort. The SMRR was 2.24 (95% CI 2.15-2.34) in males and 3.09 (95% CI 3.00-3.19) in females over the study period. SMRR decreased since 2000 to 1.59 (95% CI 1.39-1.81) for the period 2011-2015. Median survival was 26.80 years (95% CI 26.30-27.30), where age and comorbidity independently increased the risk of death. The leading causes of deaths were cardiovascular diseases (26.60%), cancer (16.80%), rheumatic diseases (5.80%), chronic pulmonary disease 491 (5.50%), dementia (3.00%), and diabetes 235 (2.6%).

CONCLUSIONS

The mortality rate in patients with RA in WA has decreased but remains 1.59-times higher than in community counterparts, suggesting that there is room for further improvement. Comorbidity is the main modifiable risk factor to further reduce mortality in patients with RA.

Determination of the main causes, outcome, and prognostic factors of patients with rheumatologic diseases admitted to the medical intensive care unit in Southern Iran

BACKGROUND

Systemic rheumatic diseases (SRD) are a heterogeneous group of diseases that can involve several organ systems and occasionally requires intensive care unit (ICU) admission because of severe systemic disease, life-threatening organ involvement, or complication of treatment. The objective of this study is to determine the causes, outcome, and prognostic factors of patients with rheumatologic diseases admitted in teaching medical ICUs in southern Iran.

METHODS

A retrospective case review of all patients with rheumatologic diseases admitted in the academic medical ICUs in two referral hospitals in southern Iran, from March 2015 to January 2020. Patients' data were documented from their hospital records and the cause of admission, in-hospital outcome, and prognostic factors was evaluated.

RESULTS

Ninety-one patients were included, of which 71.4% were female. Systemic lupus erythematosus (54.9%) was the most common disease. Nineteen (20.9%) patients were new cases of rheumatological disease. The most frequent symptom for admittance was dyspnea (54.9%) and hemoptysis (20.9%). The in-hospital mortality rate was 48%, and the leading cause of death was infection (29 patients; 65.5%) followed by disease activity (18 patients; 40.9%). Also, the death of 29.5% of patients was presumed due to both disease activity and infection. Factors associated with mortality included renal insufficiency (p < 0.028), infection (p < 0.001), pneumonia (p < 0.042), dyspnea (p < 0.042), loss of consciousness (p < 0.046), azathioprine consumption (p < 0.004) during 1 month before ICU admission, mechanical ventilation (p < 0.001), renal replacement therapy (p < 0.001), CNS involvement (p < 0.009), and ICU medications such as cyclosporine and azathioprine (0.03 and 0.03, respectively) or treatments such as plasmapheresis (p < 0.018).

CONCLUSION

The ICU mortality rate of patients with SRD was high. Infection and disease exacerbation are the leading reasons for ICU admission in systemic rheumatic diseases. Intensivists must keep in mind that SRD exacerbation may require immunosuppressive agents along with lifesaving interventions, more particularly in newly diagnosed SRDs. Key Points • The ICU mortality rate of patients with SRD was high. • Infection and disease exacerbation are the leading reasons for ICU admission in systemic rheumatic diseases. • 63.8% of our patients fall into this category of new cases of rheumatologic disease and disease flare-up.

Mortality and its risk factors in critically ill patients with connective tissue diseases: A meta-analysis

BACKGROUND

Systemic lupus erythematosus (SLE), primary Sjögren's syndrome (pSS), systemic sclerosis (SSc), idiopathic inflammatory myopathies (IIM) and rheumatoid arthritis (RA) are connective tissue diseases (CTD) whose complications can lead to management in the intensive care unit (ICU).

OBJECTIVES

To estimate by meta-analysis ICU mortality rates for CTD.

METHODS

A systematic literature review was performed to identify articles studying critically ill CTD patients. A random-effects model was chosen for analysis. Pooled proportion mortality was calculated using aggregated-data meta-analysis with a random-effects model and assessment of heterogeneity with the I² statistic. Risk of bias was assessed using the quality assessment tool.

RESULTS

Of the 5694 individual publications, a sample of 31 independent cohorts was used for the meta-analysis totalling 5007 patients. The main cause for admission was sepsis (43%) followed by "flare-ups" (40%). The overall pooled proportion of mortality of CTD patients across all 31 studies was 33% (95%CI: 28-38%). In the IIM subgroup and that of SSc, mortality was 70% (95%CI: 46-86%) and 40% (95%CI: 25-47%), respectively. In the SLE subgroup, mortality was similar to the overall pooled mortality of 35% (95%CI: 29-42%). Subgroup mortality for RA and pSS patients was respectively 20% (95%CI: 11-33%) and 17% (95%CI: 6-41%); lower than the overall pooled mortality. Heterogeneity in each subgroup remained high.

CONCLUSION

The overall pooled proportion of mortality of ICU patients with CTD was 33% (95%CI: 28-38%), with a high heterogeneity (I²= 89%). In the subgroup analysis, mortality was higher for patients with IIM and SSc.

Factors affecting patient satisfaction related to cost and treatment effectiveness in rheumatoid arthritis: results from the multicenter observational cohort study, FRANK Registry

Background

To further improve rheumatoid arthritis (RA) treatment, it is necessary to understand each RA patient’s satisfaction and to identify the factors affecting their satisfaction. Despite the rise in medical costs for RA, little is known about the factors that influence patient satisfaction with the cost of treatment in RA patients.

Methods

This is a multicenter observational study of Japanese RA patients from the FRANK Registry with data analyzed from March 2017 to August 2020. We collected data on demographic characteristics, clinical data, quality of life which was evaluated using the EuroQol 5-dimensional questionnaire (EQ5D), and patient satisfaction. The four categories of patient satisfaction were evaluated individually (i.e., cost, treatment efficacy, activities of daily living [ADL], and global treatment satisfaction). We analyzed the factors that affected each patient’s satisfaction, such as age, sex, EQ5D, disease duration, disease activity, and treatment.

Results

This study included 2235 RA outpatients (406 males, 1829 females). In RA patients, “very satisfied” and “satisfied” were given for nearly half of each satisfaction aspect (cost 49%; efficacy 72%; ADL 58%; global treatment 66%) at the time of the initial registration. To investigate the factors influencing each satisfaction, multivariate analysis has revealed that the use of b/tsDMARDs increased satisfaction of treatment effect (odds ratio [OR] 0.66) and ADL (OR 0.78) but decreased cost satisfaction (OR 2.21). Age (50–64 years; OR 0.91; 65–74 years, 0.55: ≥ 75 years, 0.35), female (OR 0.81), and history of musculoskeletal surgery (OR 0.60) all increased cost satisfaction. Patients with lower disease activity and higher EQ5D scores had higher levels of satisfaction in all areas.

Conclusions

In this study, patient satisfaction in terms of cost, treatment effect, ADL, and overall treatment was generally higher, but some patients were dissatisfied. The cost of satisfaction increased with age and a history of musculoskeletal surgery, while it decreased with a lower EQ5D score and the use of b/tsDMARDs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13075-022-02746-5.