Background The effectiveness of ultrasound-guided pericapsular nerve group (PENG) blocks for postoperative analgesia following direct anterior total hip arthroplasty under general anaesthesia remains uncertain. We conducted a prospective randomized trial to determine whether a PENG block would reduce postoperative intravenous morphine consumption compared with surgical local anaesthetic infiltration. Methods Consecutive patients were randomized to receive either an ultrasound-guided PENG block after induction of anaesthesia or surgical local anaesthetic infiltration after femoral neck resection. General anaesthesia and perioperative analgesic management were standardized according to institutional practice, and all procedures were performed by a single surgeon. The primary outcome was total intravenous morphine consumption during the post-anaesthesia care unit (PACU) stay. The secondary outcome measure was total oral morphine consumption from PACU discharge to postoperative day 1 (POD#1), expressed in oral morphine equivalents (OME). Results Of the 60 recruited patients, 56 were analysed using a modified intention-to-treat analysis. Median (interquartile range) PACU morphine consumption did not differ significantly between the control and PENG groups (0 [0–6] mg vs. 0 [0–3] mg; median difference 0 mg, 95% confidence interval (CI) −2 to 4.5; P = 0.235). Total OME consumption through POD#1 was lower in the PENG group compared to the control group (0 [0–10] mg vs. 7 [3–20] mg; P = 0.040). Conclusions In patients undergoing direct anterior total hip arthroplasty, PENG block, when implemented as part of a multimodal analgesic pathway, did not reduce PACU intravenous morphine consumption compared with surgical local infiltration.
Background Thyroid hormone disturbances are a hallmark of critical illness and are commonly attributed to non-thyroidal illness syndrome. The clinical relevance of thyroid hormone supplementation in the intensive care unit (ICU) remains uncertain. This study aimed to characterize thyroid hormone concentrations in relation to disease severity in critically ill patients receiving thyroid hormone supplementation. Methods In this single-center retrospective study, data from 98 ICU patients receiving thyroid hormone supplementation were analyzed. A total of 186 measurements of thyroid-stimulating hormone (TSH), free triiodothyronine (fT3), and free thyroxine (fT4) were evaluated in relation to disease severity, organ dysfunction, and clinical interventions. Results Lower fT3 and fT4 concentrations were consistently associated with markers of critical illness, including shock, vasopressor use, mechanical ventilation, and higher SOFA scores. TSH showed weaker and less consistent associations but correlated positively with duration of ICU stay. Thyroid hormone concentrations were largely unrelated to most therapeutic interventions, with only weak correlations observed. Notably, hormone concentrations did not differ meaningfully between patients with and without preexisting thyroid disease, apart from higher TSH concentrations and levothyroxine dosing in the former group. Conclusions Thyroid hormone concentrations in critical illness were closely associated with disease severity. As all patients received supplementation and no comparator group was available, the independent effect of treatment could not be assessed. Prospective controlled studies are needed.
Cytomegalovirus (CMV) and human herpesvirus-6 (HHV-6) infections typically follow a benign course in immunocompetent individuals. However, critically ill patients may develop profound immune dysregulation associated with sepsis, trauma, prolonged mechanical ventilation, or immunomodulatory therapy, which may predispose them to reactivation of latent viruses. A growing body of evidence suggests that CMV reactivation in immunocompetent, critically ill patients is associated with adverse clinical outcomes, including prolonged mechanical ventilation, longer ICU stay, and an increased risk of secondary infections. The clinical significance of herpesvirus reactivation in this population remains incompletely defined, and diagnosis may be challenging due to nonspecific clinical presentation and limited availability of molecular diagnostics. The aim of this study was to present a case series of CMV and HHV-6 reactivation in immunocompetent, critically ill patients and to discuss diagnostic and therapeutic challenges in the context of current literature. We describe three critically ill patients who developed herpesvirus reactivation during hospitalisation. One patient presented with clinically significant CMV reactivation suspected to contribute to gastrointestinal and pulmonary manifestations in the setting of rising CMV DNAemia. Another patient was diagnosed with CMV-associated immune thrombocytopaenia. A third patient developed rapidly progressive neurological deterioration with cerebrospinal fluid pleocytosis and HHV-6 DNA detected in cerebrospinal fluid, consistent with suspected HHV-6-associated encephalitis. Herpesvirus reactivation may be considered in critically ill patients with unexplained clinical deterioration despite the absence of classical immunosuppression.
Background Quantitative neuromuscular monitoring is essential for the safe management of neuromuscular blockade. Although electromyography (EMG) is increasingly regarded as a reference technique, kinemyography (KMG) remains widely used in clinical practice; however, clinically relevant differences between these modalities during onset and recovery of neuromuscular blockade are not fully understood. Methods In this prospective observational study, adult patients undergoing general anaesthesia with non-depolarising neuromuscular blocking agents were simultaneously monitored using EMG and KMG applied to opposite hands, and onset and recovery times of neuromuscular block were compared between modalities; additionally, a novel EMG-derived parameter (T1%) was explored as a marker of recovery. Results In the 137 analysed patients, EMG and KMG differed during onset and recovery, with greater and only statistically significant disagreement during onset (EMG 158 ± 74 s vs. KMG 118 ± 68 s, P < 0.001; recovery: EMG 27 ± 22 min vs. KMG 24 ± 21 min, P = 0.242). T1% tended to be higher with sugammadex than with spontaneous recovery (P = 0.074). Conclusions KMG and EMG are not interchangeable for perioperative neuromuscular monitoring; the observed disagreement suggests that KMG may overestimate neuromuscular block, whereas EMG-based monitoring may provide a safer approach. The EMG-derived parameter T1% appears promising but requires further validation before routine clinical use.
Introduction Large language model (LLM)-as-a-judge systems offer scalable evaluation of artificial intelligence (AI)-generated clinical outputs, yet their susceptibility to prompt variability raises concerns regarding reproducibility and alignment with expert judgement. This study examined whether evaluation prompt strategies influence scoring patterns and concordance with clinical raters in critical care. Material and methods This post-hoc analysis used 90 structured clinical reports generated in a prior study using an XGBoost ICU mortality prediction model trained on the MIMIC-IV database. GPT-4o (Azure AI, version 2024-11-20) produced structured interpretations from risk estimates and SHAP attributions. These outputs were evaluated using the IMPACT framework under three evaluation prompt strategies: baseline (E1), top-down decremental (E2), and bottom-up incremental (E3). Agreement between clinician ratings and the automated o3-mini evaluator (Azure AI, version 2025-01-31) was assessed using intraclass correlation coefficients (ICC), with strategy comparisons by Fisher’s z-transformation. Score deviations were examined with repeated-measures ANOVA. Results Mean IMPACT scores were 79.9 (SD 9.9) for E1, 83.3 (SD 9.6) for E2, 78.7 (SD 9.1) for E3, and 78.6 (SD 8.9) for clinicians. All strategies demonstrated substantial agreement (ICC > 0.80). E2 showed significantly lower agreement with clinicians (ICC = 0.82) than E1 and E3 (both ICC = 0.94, p < 0.001). Score deviations differed significantly across strategies (p < 0.001), with E3 showing the smallest mean deviation (0.1) and E2 the largest (4.7). Conclusions Prompt design meaningfully affects both IMPACT scoring patterns and the reliability of LLM-based evaluators. Bottom-up incremental scoring showed the closest alignment with human assessment, underscoring the need for standardised prompt architectures in clinical AI evaluation.
