Prognosis assessment model based on low serum calcium in patients with acute pulmonary thromboembolism

Federated-learning-based prognosis assessment model for acute pulmonary thromboembolism

BACKGROUND

Acute pulmonary thromboembolism (PTE) is a common cardiovascular disease and recognizing low prognosis risk patients with PTE accurately is significant for clinical treatment. This study evaluated the value of federated learning (FL) technology in PTE prognosis risk assessment while ensuring the security of clinical data.

METHODS

A retrospective dataset consisted of PTE patients from 12 hospitals were collected, and 19 physical indicators of patients were included to train the FL-based prognosis assessment model to predict the 30-day death event. Firstly, multiple machine learning methods based on FL were compared to choose the superior model. And then performance of models trained on the independent (IID) and non-independent identical distributed(Non-IID) datasets was calculated and they were tested further on Real-world data. Besides, the optimal model was compared with pulmonary embolism severity index (PESI), simplified PESI (sPESI), Peking Union Medical College Hospital (PUMCH).

RESULTS

The area under the receiver operating characteristic curve (AUC) of logistic regression(0.842) outperformed convolutional neural network (0.819) and multi layer perceptron (0.784). Under IID, AUC of model trained using FL(Fed) on the training, validation and test sets was 0.852 ± 0.002, 0.867 ± 0.012 and 0.829 ± 0.004. Under Real-world, AUC of Fed was 0.855 ± 0.005, 0.882 ± 0.003 and 0.835 ± 0.005. Under IID and Real-world, AUC of Fed surpassed centralization model(NonFed) (0.847 ± 0.001, 0.841 ± 0.001 and 0.811 ± 0.001). Under Non-IID, although AUC of Fed (0.846 ± 0.047) outperformed NonFed (0.841 ± 0.001) on validation set, it (0.821 ± 0.016 and 0.799 ± 0.031) slightly lagged behind NonFed (0.847 ± 0.001 and 0.811 ± 0.001) on the training and test sets. In practice, AUC of Fed (0.853, 0.884 and 0.842) outshone PESI (0.812, 0.789 and 0.791), sPESI (0.817, 0.770 and 0.786) and PUMCH(0.848, 0.814 and 0.832) on the training, validation and test sets. Additionally, Fed (0.842) exhibited higher AUC values across test sets compared to those trained directly on the clients (0.758, 0.801, 0.783, 0.741, 0.788).

CONCLUSIONS

In this study, the FL based machine learning model demonstrated commendable efficacy on PTE prognostic risk prediction, rendering it well-suited for deployment in hospitals.

The association between albumin-corrected calcium and prognosis in patients with cardiac arrest: a retrospective study based on the MIMIC-IV database

BACKGROUND

Cardiac arrest (CA) is one of the leading causes of death globally, characterized by high incidence and mortality. It is of particular significance to determine the prognosis of patients with CA early and accurately. Therefore, we aim to investigate the correlation between albumin-corrected calcium (ACC) and the prognosis in patients diagnosed with CA.

METHODS

We retrospectively collected data from medical information mart for intensive care IV database. Patients were divided into two groups (survival and non-survival groups), according to the 90-day prognosis. In the Restricted cubic spline (RCS) analysis, the cut-off values (8.86 and 10.32) were obtained to categorize patients into three groups: low ACC group (< 8.86), moderate ACC group (8.86-10.32), and high ACC group (> 10.32). The least absolute shrinkage and selection operator with a ten-fold cross-validation regression analysis was performed to identify variables linked to the mortality. The inverse probability treatment weighting (IPTW) was used to address the confounding factors, and a weighted cohort was generated. RCS, Kaplan-Meier curve, and Cox regression analyses were used to explore the relationship between ACC and the mortality. Sensitivity analysis was employed to validate the stability of the results.

RESULTS

Cut-off values for ACC of 8.86 and 10.32 were determined. RCS analyses showed that there was an overall non-linear trend relationship between ACC and the risk of 90-day and 360-day mortalities. After IPTW adjustment, compared to the moderate ACC group, the 90-day and 360-day mortalities in the high ACC group were higher (P < 0.05). The Cox analyses before and after IPTW adjustment showed that both low ACC and high ACC group were independent risk factors for 90-day and 360-day all-cause mortality in patients with CA (P < 0.05). The results obtained from sensitivity analyses indicated the stability of the findings. The Kaplan-Meier survival curves indicated that 90- and 360-day cumulative survival rates in the low ACC and high ACC groups were lower than that in the moderate ACC group (χ2 = 11.350, P = 0.003; χ2 = 14.110, P = 0.001).

CONCLUSION

Both low ACC (< 8.86) and high ACC groups (> 10.32) were independent risk factors for 90-day and 360-day all-cause mortality in patients with CA (P < 0.05). For those CA patients with high and low ACC, it deserved the attention of clinicians.

Additive prognostic value of serum calcium to the ESC risk stratification in patients with acute pulmonary embolism

BACKGROUND

Hypocalcemia has been shown to be involved in the adverse outcomes of acute pulmonary embolism (APE). We aimed to determine the incremental value of adding hypocalcemia, defined as serum calcium level ≤ 2.12 mmol/L, on top of the European Society of Cardiology (ESC) prognostic algorithm, for the prediction of in-hospital mortality in APE patients, which in turn could lead to the optimization of APE management.

METHODS

This study was conducted at West China Hospital of Sichuan University from January 2016 to December 2019. Patients with APE were retrospectively analyzed and divided into 2 groups based on serum calcium levels. Associations between hypocalcemia and adverse outcomes were assessed by Cox analysis. The accuracy of risk stratification for in-hospital mortality was assessed with the addition of serum calcium to the current ESC prognostic algorithm.

RESULTS

Among 803 patients diagnosed with APE, 338 (42.1%) patients had serum calcium levels ≤ 2.12 mmol/L. Hypocalcemia was significantly associated with higher in-hospital and 2-year all-cause mortality compared to the control group. The addition of serum calcium to ESC risk stratification enhanced net reclassification improvement. Low-risk group with serum calcium level > 2.12 mmol/L had a 0% mortality rate, improving the negative predictive value up to 100%, while high-risk group with serum calcium level ≤ 2.12 mmol/L indicated a higher mortality of 25%.

CONCLUSION

Our study identified serum calcium as a novel predictor of mortality in patients with APE. In the future, serum calcium may be added to the commonly used ESC prognostic algorithm for better risk stratification of patients suffering from APE.

Non-linear relationship between albumin-corrected calcium and 30-day in-hospital mortality in ICU patients: A multicenter retrospective cohort study

OBJECTIVE

Albumin-corrected calcium is usually calculated to reflect the real serum calcium level of the whole body by physicians. However, studies on the association between albumin-corrected calcium and 30-day in-hospital mortality in Intensive Care Unit (ICU) patients are rare. The purpose of our study was to explore the association between baseline albumin-corrected calcium and 30-day in-hospital mortality in the American ICU population.

METHODS

A multicenter retrospective cohort study of 102,245 ICU patients in the eICU-CRD v2.0 from the USA during 2014-2015 was performed. The average age was 63.7 ± 16.9 years, of which 55,313 (53.7%) were men and 47,758 (46.3%) were women. The association between albumin-corrected calcium and 30-day in-hospital mortality was analyzed by Cox proportional-hazards regression, smooth curve fitting, piecewise linear regression, subgroup analyses, and a series of sensitivity analyses.

RESULTS

We found that among ICU patients with calcium abnormalities, more than 95% were mild hypocalcemia or mild hypercalcemia. The risk of 30-day in-hospital mortality will increase by 10% in the ≥7.5-< 8.5 mg/dl subgroup (OR=1.1, 95% CI 1.0-1.3) or 20% in the ≥10.3-<12 mg/dl subgroup (OR=1.2, 95% CI 1.1-1.3) when the albumin-corrected calcium level increases by 1 mg/dl. Additionally, the relationship between albumin-corrected calcium and 30-day in-hospital mortality was U shaped; the inflection point was 8.9 mg/dl (log likelihood ratio test P = 0.005). Finally, after a series of sensitivity analyses, we found that the relationship between albumin-corrected calcium and 30-day in-hospital mortality remained significant.

CONCLUSION

In a large nationally representative cohort of ICU patients, abnormalities in albumin-corrected calcium, particularly slight hypocalcemia or slight hypercalcemia, were associated with an increased 30-day in-hospital mortality risk, and yet the findings in this study need to be further confirmed by prospective studies.