Publications

BACKGROUND

His-bundle pacing (HBP) has emerged as an alternative to conventional ventricular pacing because of its ability to deliver physiological ventricular activation. Pacing at the His bundle produces different electrocardiographic (ECG) responses: selective His-bundle pacing (S-HBP), non-selective His bundle pacing (NS-HBP), and myocardium-only capture (MOC). These 3 capture types must be distinguished from each other, which can be challenging and time-consuming even for experts.

OBJECTIVE

The purpose of this study was to use artificial intelligence (AI) in the form of supervised machine learning using a convolutional neural network (CNN) to automate HBP ECG interpretation.

METHODS

We identified patients who had undergone HBP and extracted raw 12-lead ECG data during S-HBP, NS-HBP, and MOC. A CNN was trained, using 3-fold cross-validation, on 75% of the segmented QRS complexes labeled with their capture type. The remaining 25% was kept aside as a testing dataset.

RESULTS

The CNN was trained with 1297 QRS complexes from 59 patients. Cohen kappa for the neural network's performance on the 17-patient testing set was 0.59 (95% confidence interval 0.30 to 0.88; <.0001), with an overall accuracy of 75%. The CNN's accuracy in the 17-patient testing set was 67% for S-HBP, 71% for NS-HBP, and 84% for MOC.

CONCLUSION

We demonstrated proof of concept that a neural network can be trained to automate discrimination between HBP ECG responses. When a larger dataset is trained to higher accuracy, automated AI ECG analysis could facilitate HBP implantation and follow-up and prevent complications resulting from incorrect HBP ECG analysis.

INTRODUCTION

Liver disease is a common cause of non-AIDS morbidity and mortality in people living with HIV (PLHIV), but the prevalence and significance of liver function test (LFT) abnormalities in early HIV infection is unknown. This study aimed to characterize LFTs in a large cohort of participants with acute HIV infection initiating immediate antiretroviral therapy (ART) and examine the association between LFTs and biomarkers of HIV infection and inflammation.

METHODS

We measured LFTs at the time of HIV diagnosis and at 4, 12, 24 and 48 weeks after ART initiation in 426 Thai individuals with acute HIV infection from 2009 to 2018. A subset of individuals had data available at 96 and 144 weeks. We excluded individuals with concomitant viral hepatitis. Alanine aminotransferase (ALT) was the primary outcome of interest; values greater than 1.25 times the upper limit of normal were considered elevated. Analyses utilized descriptive statistics, non-parametric tests and multivariate logistic regression.

RESULTS

Sixty-six of the 426 individuals (15.5%) had abnormal baseline ALT levels; the majority (43/66, 65.5%) had Grade 1 elevations. Elevated baseline ALT correlated with Fiebig stages III to V (p = 0.001) and baseline HIV RNA >6 log copies/mL (p = 0.012). Baseline elevations resolved by 48 weeks on ART in 59 of the 66 individuals (89%). ALT elevations at 24 and 48 weeks correlated with Fiebig stages I to II at diagnosis (p < 0.001), baseline plasma HIV RNA levels <6 log copies/mL (p < 0.001), abnormal baseline ALT (p < 0.001), baseline CD4 >350 cells/μL (p = 0.03) and older age (p = 0.03). Individuals initiating efavirenz-based regimens were more likely to have elevated ALT levels at 48 weeks compared with those on non-efavirenz-based regimens (p = 0.003).

CONCLUSIONS

One in six people with acute HIV infection have elevated LFTs. Clinical outcomes with ART started in acute HIV are generally good, with resolution of ALT elevations within 48 weeks on ART in most cases. These results suggest a multifactorial model for hepatic injury involving a combination of HIV-associated and ART-associated processes, which may change over time.