Clinical benefit of AI-assisted lung ultrasound in a resource-limited intensive care unit.

BACKGROUND: Interpreting point-of-care lung ultrasound (LUS) images from intensive care unit (ICU) patients can be challenging, especially in low- and middle- income countries (LMICs) where there is limited training available. Despite recent advances in the use of Artificial Intelligence (AI) to automate many ultrasound imaging analysis tasks, no AI-enabled LUS solutions have been proven to be clinically useful in ICUs, and specifically in LMICs. Therefore, we developed an AI solution that assists LUS practitioners and assessed its usefulness in  a low resource ICU. METHODS: This was a three-phase prospective study. In the first phase, the performance of four different clinical user groups in interpreting LUS clips was assessed. In the second phase, the performance of 57 non-expert clinicians with and without the aid of a bespoke AI tool for LUS interpretation was assessed in retrospective offline clips. In the third phase, we conducted a prospective study in the ICU where 14 clinicians were asked to carry out LUS examinations in 7 patients with and without our AI tool and we interviewed the clinicians regarding the usability of the AI tool. RESULTS: The average accuracy of beginners' LUS interpretation was 68.7% [95% CI 66.8-70.7%] compared to 72.2% [95% CI 70.0-75.6%] in intermediate, and 73.4% [95% CI 62.2-87.8%] in advanced users. Experts had an average accuracy of 95.0% [95% CI 88.2-100.0%], which was significantly better than beginners, intermediate and advanced users (p < 0.001). When supported by our AI tool for interpreting retrospectively acquired clips, the non-expert clinicians improved their performance from an average of 68.9% [95% CI 65.6-73.9%] to 82.9% [95% CI 79.1-86.7%], (p < 0.001). In prospective real-time testing, non-expert clinicians improved their baseline performance from 68.1% [95% CI 57.9-78.2%] to 93.4% [95% CI 89.0-97.8%], (p < 0.001) when using our AI tool. The time-to-interpret clips improved from a median of 12.1 s (IQR 8.5-20.6) to 5.0 s (IQR 3.5-8.8), (p < 0.001) and clinicians' median confidence level improved from 3 out of 4 to 4 out of 4 when using our AI tool. CONCLUSIONS: AI-assisted LUS can help non-expert clinicians in an LMIC ICU improve their performance in interpreting LUS features more accurately, more quickly and more confidently.

Heart Rate Variability Measured from Wearable Devices as a Marker of Disease Severity in Tetanus.

Tetanus is a disease associated with significant morbidity and mortality. Heart rate variability (HRV) is an objective clinical marker with potential value in tetanus. This study aimed to investigate the use of wearable devices to collect HRV data and the relationship between HRV and tetanus severity. Data were collected from 110 patients admitted to the intensive care unit in a tertiary hospital in Vietnam. HRV indices were calculated from 5-minute segments of 24-hour electrocardiogram recordings collected using wearable devices. HRV was found to be inversely related to disease severity. The standard deviation of NN intervals and interquartile range of RR intervals (IRRR) were significantly associated with the presence of muscle spasms; low frequency (LF) and high frequency (HF) indices were significantly associated with severe respiratory compromise; and the standard deviation of differences between adjacent NN intervals, root mean square of successive differences between normal heartbeats, LF to HF ratio, total frequency power, and IRRR, were significantly associated with autonomic nervous system dysfunction. The findings support the potential value of HRV as a marker for tetanus severity, identifying specific indices associated with clinical severity thresholds. Data were recorded using wearable devices, demonstrating this approach in resource-limited settings where most tetanus occurs.

Urinary catecholamine excretion, cardiovascular variability, and outcomes in tetanus.

Severe tetanus is characterized by muscle spasm and cardiovascular system disturbance. The pathophysiology of muscle spasm is relatively well understood and involves inhibition of central inhibitory synapses by tetanus toxin. That of cardiovascular disturbance is less clear, but is believed to relate to disinhibition of the autonomic nervous system. The clinical syndrome of autonomic nervous system dysfunction (ANSD) seen in severe tetanus is characterized principally by changes in heart rate and blood pressure which have been linked to increased circulating catecholamines. Previous studies have described varying relationships between catecholamines and signs of ANSD in tetanus, but are limited by confounders and assays used. In this study, we aimed to perform detailed characterization of the relationship between catecholamines (adrenaline and noradrenaline), cardiovascular parameters (heart rate and blood pressure) and clinical outcomes (ANSD, mechanical ventilation required, and length of intensive care unit stay) in adults with tetanus, as well as examine whether intrathecal antitoxin administration affected subsequent catecholamine excretion. Noradrenaline and adrenaline were measured by ELISA from 24-h urine collections taken on day 5 of hospitalization in 272 patients enrolled in a 2 × 2 factorial-blinded randomized controlled trial in a Vietnamese hospital. Catecholamine results measured from 263 patients were available for analysis. After adjustment for potential confounders (i.e., age, sex, intervention treatment, and medications), there were indications of non-linear relationships between urinary catecholamines and heart rate. Adrenaline and noradrenaline were associated with subsequent development of ANSD, and length of ICU stay.

Human versus equine intramuscular antitoxin, with or without human intrathecal antitoxin, for the treatment of adults with tetanus: a 2 × 2 factorial randomised controlled trial.

BACKGROUND: Intramuscular antitoxin is recommended in tetanus treatment, but there are few data comparing human and equine preparations. Tetanus toxin acts within the CNS, where there is limited penetration of peripherally administered antitoxin; thus, intrathecal antitoxin administration might improve clinical outcomes compared with intramuscular injection. METHODS: In a 2  × 2 factorial trial, all patients aged 16 years or older with a clinical diagnosis of generalised tetanus admitted to the intensive care unit of the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam, were eligible for study entry. Participants were randomly assigned first to 3000 IU human or 21 000 U equine intramuscular antitoxin, then to either 500 IU intrathecal human antitoxin or sham procedure. Interventions were delivered by independent clinicians, with attending clinicians and study staff masked to treatment allocations. The primary outcome was requirement for mechanical ventilation. The analysis was done in the intention-to-treat population. The study is registered at ClinicalTrials.gov, NCT02999815; recruitment is completed. FINDINGS: 272 adults were randomly assigned to interventions between Jan 8, 2017, and Sept 29, 2019, and followed up until May, 2020. In the intrathecal allocation, 136 individuals were randomly assigned to sham procedure and 136 to antitoxin; in the intramuscular allocation, 109 individuals were randomly assigned to equine antitoxin and 109 to human antitoxin. 54 patients received antitoxin at a previous hospital, excluding them from the intramuscular antitoxin groups. Mechanical ventilation was given to 56 (43%) of 130 patients allocated to intrathecal antitoxin and 65 (50%) of 131 allocated to sham procedure (relative risk [RR] 0·87, 95% CI 0·66-1·13; p=0·29). For the intramuscular allocation, 48 (45%) of 107 patients allocated to human antitoxin received mechanical ventilation compared with 48 (44%) of 108 patients allocated to equine antitoxin (RR 1·01, 95% CI 0·75-1·36, p=0·95). No clinically relevant difference in adverse events was reported. 22 (16%) of 136 individuals allocated to the intrathecal group and 22 (11%) of 136 allocated to the sham procedure experienced adverse events related or possibly related to the intervention. 16 (15%) of 108 individuals allocated to equine intramuscular antitoxin and 17 (16%) of 109 allocated to human antitoxin experienced adverse events related or possibly related to the intervention. There were no intervention-related deaths. INTERPRETATION: We found no advantage of intramuscular human antitoxin over intramuscular equine antitoxin in tetanus treatment. Intrathecal antitoxin administration was safe, but did not provide overall benefit in addition to intramuscular antitoxin administration. FUNDING: The Wellcome Trust.

The Role of the Gastrointestinal Tract in Toxigenic Clostridium tetani Infection: A Case-Control Study.

Tetanus arises from wound contamination with Clostridium tetani, but approximately one fifth of patients have no discernable entry wound. Clostridium tetani is culturable from animal feces, suggesting the gastrointestinal tract could be an endogenous reservoir or direct-entry portal, but human data are lacking. In this study of 101 Vietnamese adults with tetanus and 29 hospitalized control subjects, admission stool samples were cultured for C. tetani. Anti-tetanus toxin antibodies were measured by ELISA. Clostridium tetani toxigenicity was evaluated using polymerase chain reaction and sequencing. Toxigenic C. tetani was cultured from stool samples in 50 of 100 (50%) tetanus cases and 12 of 28 (42.9%) control subjects (P = 0.50), and stool samples of 44 of 85 (52.4%) tetanus cases with clinically identified wounds compared with 6 of 15 (47.6%) patients without clinically identified wounds (P = 0.28). Nine of 12 (75%) control subjects with toxigenic C. tetani in their stool samples lacked protective antibody concentrations. These findings fail to show evidence of an association between gastrointestinal C. tetani and tetanus infection, but emphasize the importance of increasing vaccination coverage.

Severity detection tool for patients with infectious disease.

Hand foot and mouth disease (HFMD) and tetanus are serious infectious diseases in low- and middle-income countries. Tetanus, in particular, has a high mortality rate and its treatment is resource-demanding. Furthermore, HFMD often affects a large number of infants and young children. As a result, its treatment consumes enormous healthcare resources, especially when outbreaks occur. Autonomic nervous system dysfunction (ANSD) is the main cause of death for both HFMD and tetanus patients. However, early detection of ANSD is a difficult and challenging problem. The authors aim to provide a proof-of-principle to detect the ANSD level automatically by applying machine learning techniques to physiological patient data, such as electrocardiogram waveforms, which can be collected using low-cost wearable sensors. Efficient features are extracted that encode variations in the waveforms in the time and frequency domains. The proposed approach is validated on multiple datasets of HFMD and tetanus patients in Vietnam. Results show that encouraging performance is achieved. Moreover, the proposed features are simple, more generalisable and outperformed the standard heart rate variability analysis. The proposed approach would facilitate both the diagnosis and treatment of infectious diseases in low- and middle-income countries, and thereby improve patient care.

Heart Rate Variability as an Indicator of Autonomic Nervous System Disturbance in Tetanus.

Autonomic nervous system dysfunction (ANSD) is a significant cause of mortality in tetanus. Currently, diagnosis relies on nonspecific clinical signs. Heart rate variability (HRV) may indicate underlying autonomic nervous system activity and represents a potentially valuable noninvasive tool for ANSD diagnosis in tetanus. HRV was measured from three 5-minute electrocardiogram recordings during a 24-hour period in a cohort of patients with severe tetanus, all receiving mechanical ventilation. HRV measurements from all subjects-five with ANSD (Ablett Grade 4) and four patients without ANSD (Ablett Grade 3)-showed HRV was lower than reported ranges for healthy individuals. Comparing different severities of tetanus, raw data for both time and frequency measurements of HRV were reduced in those with ANSD compared with those without. Differences were statistically significant in all except root mean square SD, indicating HRV may be a valuable tool in ANSD diagnosis.

A Pilot Study to Assess Safety and Feasibility of Intrathecal Immunoglobulin for the Treatment of Adults with Tetanus.

Tetanus remains a significant burden in many low- and middle-income countries. The tetanus toxin acts within the central nervous system and intrathecal antitoxin administration may be beneficial, but there are safety concerns, especially in resource-limited settings. We performed a pilot study to assess the safety and feasibility of intrathecal human tetanus immunoglobulin in five adults with tetanus before the conduct of a large randomized controlled trial. Intrathecal injection via lumbar puncture was given to all patients within a median 140 (range 100-165) minutes of intensive care unit (ICU) admission. There were no serious adverse effects associated with the procedure although three patients had probably related minor adverse events which resolved spontaneously. Median ICU length of stay was 14 (range 5-17) days. Two patients required mechanical ventilation and one developed a deep vein thrombosis. Within 240 days of hospital discharge, no patients died and all patients returned to work.

A one-year prospective study of colonization with antimicrobial-resistant organisms on admission to a Vietnamese intensive care unit.

There is a paucity of data regarding initial bacterial colonization on admission to Intensive Care Units (ICUs) in low and middle-income countries (LMICs). Patients admitted to ICUs in LMICs are at high-risk of subsequent infection with antimicrobial-resistant organisms (AROs). We conducted a prospective, observational study at the Hospital for Tropical Diseases in Ho Chi Minh City, Vietnam from November 2014 to January 2016 to assess the colonization and antimicrobial susceptibility of Staphylococcus aureus, Escherichia coli, Klebsiella spp., Pseudomonas spp. and Acinetobacter spp. among adult patients within 48 hours of ICU admission. We found the admission colonization prevalence (with at least one of the identified organisms) was 93.7% (785/838) and that of AROs was 63.1% (529/838). The colonization frequency with AROs among patients admitted from the community was comparable to those transferred from other hospitals (62.2% vs 63.8%). Staphylococcus aureus was the most commonly isolated bacteria from nasal swabs (13.1%, 110/838) and the methicillin-resistant Staphylococcus aureus nasal colonization prevalence was 8.6% (72/838). We isolated Escherichia coli from rectal swabs from almost all enrolled patients (88.3%, 740/838) and 52.1% (437/838) of patients were colonized by extended spectrum ß-lactamase producing Escherichia coli. Notably, Klebsiella pneumoniae was the most frequently isolated bacteria from the tracheal swabs (11.8%, 18/153). Vietnamese ICU patients have a high rate of colonization with AROs and are thus at risk of subsequent infections with these organisms if good infection control practices are not in place.