Lukáš Lafférs

Background and aims: Alcohol-associated hepatitis (AH) frequently triggers acute decompensation (AD) in cirrhosis, with severe AH linked to high short-term mortality, especially in acute-on-chronic liver failure. Current corticosteroid treatments have limited efficacy, highlighting the need for new therapies. We hypothesized that severe AH outcomes are influenced by early specialized care; thus, we examined the impact of time-to-tertiary care (TTTc). Methods: Adults with cirrhosis or advanced chronic liver disease were enrolled (RH7, NCT04767945). AH was diagnosed using National Institute on Alcohol Abuse and Alcoholism criteria. Primary admission site, TTTc, and adverse outcomes (death or liver transplantation) were analyzed. Patients admitted directly to tertiary care were assigned a TTTc of zero. Results: Of 221 AD-AH patients, 107 were transferred from secondary care to tertiary care (TTTc >0) and 114 were admitted directly (TTTc = 0). TTTc >0 patients were younger (48.3 vs. 52 years, P = .008) and had more severe disease, as shown by model for end-stage liver disease scores (25.5 vs. 20.8, P < .001) and Maddrey’s discriminant function (59.3 vs. 40.6, P < .001). Propensity-score matching yielded 49 case pairs. The Cox model showed that transfer from secondary care was not associated with increased risk, but delayed transfer (days, hazard ratio = 1.03, 95% confidence interval 1.01–1.05) independently predicted adverse outcomes. Conclusions: Delayed initiation of specialized care adversely impacts outcomes in AD-AH. If validated, timely care bundles could improve AH survival, similar to sepsis or vascular syndromes. Highlights: AD-AH is a common syndrome associated with high short-term mortality. There is an unmet need for new prognosis-modifying therapies for AH. Currently, in real-life hepatology, refining the existing bundle of care is the only practical option to improve the prognosis of AD-AH. Past experience with acute coronary syndromes, stroke, and sepsis, emphasizing symptoms-to-intervention duration, combined with the recent COVID-19 lockdown finding of increased mortality due to skewed access to specialized liver care indicates that focusing on timely specialized care might be key to improved outcome in certain liver conditions. In this line, we set out to track the number of days elapsing between admission to SC and referral to TC, coining this interval as “time-to-tertiary care” (TTTc). We examined TTTc as a potential compound surrogate that might influence the prognosis in AD-AH. After correcting for important baseline differences, we conclude that the delay of transfer to the tertiary care hospital was independently associated with a worse prognosis with each additional day in TTTc increasing adverse outcomes by nearly 3%.

Rationale: Severe alcohol-associated hepatitis (SAH) is the most critical, acute, inflammatory phenotype within the alcohol-associated liver disease (ALD) spectrum, characterized by high 30- and 90-day mortality. Since several decades, corticosteroids (CS) are the only approved pharmacotherapy offering highly limited survival benefits. Contextually, there is an evident demand for 3PM innovation in the area meeting patients’ needs and improving individual outcomes. Fecal microbiota transplantation (FMT) has emerged as one of the new potential therapeutic options. In this study, we aimed to address the crucial 3PM domains in order to assess (i) the impact of FMT on mortality in SAH patients beyond CS, (ii) to identify factors associated with the outcome to be improved (iii) the prediction of futility, (iv) prevention of suboptimal individual outcomes linked to increased mortality, and (v) personalized allocation of therapy. Methods: We conducted a prospective study (NCT04758806) in adult patients with SAH who were non-responders (NR) to or non-eligible (NE) for CS between January 2018 and August 2022. The intervention consisted of five 100 ml of FMT, prepared from 30 g stool from an unrelated healthy donor and frozen at − 80 °C, administered daily to the upper gastrointestinal (GI) tract. We evaluated the impact of FMT on 30- and 90-day mortality which we compared to the control group selected by the propensity score matching and treated by the standard of care; the control group was derived from the RH7 registry of patients hospitalized at the liver unit (NCT04767945). We have also scrutinized the FMT outcome against established and potential prognostic factors for SAH — such as the model for end-stage liver disease (MELD), Maddrey Discriminant Function (MDF), acute-on-chronic liver failure (ACLF), Liver Frailty Index (LFI), hepatic venous-portal pressure gradient (HVPG) and Alcoholic Hepatitis Histologic Score (AHHS) — to see if the 3PM method assigns them a new dimension in predicting response to therapy, prevention of suboptimal individual outcomes, and personalized patient management. Results: We enrolled 44 patients with SAH (NR or NE) on an intention-to-treat basis; we analyzed 33 patients per protocol for associated factors (after an additional 11 being excluded for receiving less than 5 doses of FMT), and 31 patients by propensity score matching for corresponding individual outcomes, respectively. The mean age was 49.6 years, 11 patients (33.3%) were females. The median MELD score was 29, and ACLF of any degree had 27 patients (81.8%). FMT improved 30-day mortality (p = 0.0204) and non-significantly improved 90-day mortality (p = 0.4386). Univariate analysis identified MELD ≥ 30, MDF ≥ 90, and ACLF grade > 1 as significant predictors of 30-day mortality, (p = 0.031; p = 0.014; p = 0.034). Survival was not associated with baseline LFI, HVPG, or AHHS. Conclusions and recommendations in the framework of 3PM: In the most difficult-to-treat sub-cohort of patients with SAH (i.e., NR/NE), FMT improved 30-day mortality. Factors associated with benefit included MELD ≤ 30, MDF ≤ 90, and ACLF < 2. These results support the potential of gut microbiome as a therapeutic target in the context of 3PM research and vice versa — to use 3PM methodology as the expedient unifying template for microbiome research. The results allow for immediate impact on the innovative concepts of (i) personalized phenotyping and stratification of the disease for the clinical research and practice, (ii) multilevel predictive diagnosis related to personalized/precise treatment allocation including evidence-based (ii) prevention of futile and sub-optimally effective therapy, as well as (iii) targeted prevention of poor individual outcomes in patients with SAH. Moreover, our results add to the existing evidence with the potential to generate new research along the SAH’s pathogenetic pathways such as diverse individual susceptibility to alcohol toxicity, host-specific mitochondrial function and systemic inflammation, and the role of gut dysbiosis thereof.

Causal mediation analysis aims at disentangling a treatment effect into an indirect mechanism operating through an intermediate outcome or mediator, as well as the direct effect of the treatment on the outcome of interest. However, the evaluation of direct and indirect effects is frequently complicated by non-ignorable selection into the treatment and/or mediator, even after controlling for observables, as well as sample selection/outcome attrition. We propose a method for bounding direct and indirect effects in the presence of such complications using a method that is based on a sequence of linear programming problems. Considering inverse probability weighting by propensity scores, we compute the weights that would yield identification in the absence of complications and perturb them by an entropy parameter reflecting a specific amount of propensity score misspecification to set-identify the effects of interest. We apply our method to data from the National Longitudinal Survey of Youth 1979 to derive bounds on the explained and unexplained components of a gender wage gap decomposition that is likely prone to non-ignorable mediator selection and outcome attrition.

Summary: Introduction: Slovakia is a country with the highest prevalence of liver cirrhosis in the world and a country with the highest proportion of Roma ethnicity at the same time. However, there is only little evidence of Roma representation in national cohorts with cirrhosis. Aims: 1. To determine the prevalence of Roma ethnicity in our cirrhosis and liver transplant registers; to compare their 2. fundamental characteristics and 3. final results with patients from the majority population. Patients and methods: A retrospective study; we acquired data from 1. Cirrhosis registry RH7; 2. Liver transplant registry: a) patients listed active on the liver-transplant waiting list; b) patients underwent first LT. The first source – the cirrhosis registry RH7 (NCT 04767945; since 2014, RH7 has been listing consecutive patients admitted to hospital with liver cirrhosis). Up to 2021, the mode of the ethnicity determination was so-called “ascribed ethnicity”. The second source – the Liver transplant registry (since 2008); the mode of ethnicity determination was identical to the one of RH7. Apart from the ethnicity, the following points were recorded and analyzed in all patients: demographics, elementary cirrhosis-relevant clinical variables such as etiology and MELD score, as well as an elementary LT-relevant variables, such as waiting time and mortality. Results: We present the results on Roma ethnic group in three cohorts from two datasets, i.e. on 1,515 patients from RH7, on 464 waitlisted patients from LT registry and on 302 transplanted patients from LT registry, respectively. The representation of Roma ethnicity in these cohorts were 2%, 4%, and 4%, respectively. Significant differences in age and gender were detected in Roma cirrhotic patients: 46 vs. 55 years (P = 0.001) and female gender 25% vs. 39% (P = 0.042). Of the first time waitlisted candidates for LT, Roma patients were also significantly younger – 42.6 vs. 51.5 years; in addition, Romas had a less prevalent alcohol-associated etiology (ALD) and a more prevalent autoimmune etiology. Finally, Roma patients after first LT were younger – 40.2 vs. 51.6 years, again with lower etiology of ALD – 15% vs. 47% and more autoimmune etiology – 39% vs. 23%. The results of Romas from all cohorts in tertiary care were comparable. Conclusion: 1. the admission of Romas to a tertiary liver care is lower than expected, for unknown reasons; 2. the age of Romas entering tertiary care is approximately ten years lower; 3. the results of Romas in tertiary care is comparable to the majority population. Key words: liver cirrhosis – Roma ethnicity – tertiary liver care – waiting list – liver transplantation

Summary: We consider evaluating the causal effects of dynamic treatments, i.e., of mul-tiple treatment sequences in various periods, based on double machine learning to control for observed, time-varying covariates in a data-driven way under a selection-on-observables assumption. To this end, we make use of so-called Neyman-orthogonal score functions, which imply the robustness of treatment effect estimation to moderate (local) misspecifications of the dynamic outcome and treatment models. This robustness property permits approximating outcome and treatment models by double machine learning even under high-dimensional covariates. In addition to effect estimation for the total population, we consider weighted estimation that permits assessing dynamic treatment effects in specific subgroups, e.g., among those treated in the first treatment period. We demonstrate that the estimators are asymptotically normal and $\sqrt{n}$-consistent under specific regularity conditions and investigate their finite sample properties in a simulation study. Finally, we apply the methods to the Job Corps study.

This paper combines causal mediation analysis with double machine learning for a data-driven control of observed confounders in a high-dimensional setting. The average indirect effect of a binary treatment and the unmediated direct effect are estimated based on efficient score functions, which are robust w.r.t. misspecifications of the outcome, mediator, and treatment models. This property is key for selecting these models by double machine learning, which is combined with data splitting to prevent overfitting. We demonstrate that the effect estimators are asymptotically normal and n^−1/2-consistent under specific regularity conditions and investigate the finite sample properties of the suggested methods in a simulation study when considering lasso as machine learner. We also provide an empirical application to the U.S. National Longitudinal Survey of Youth, assessing the indirect effect of health insurance coverage on general health operating via routine checkups as mediator, as well as the direct effect.

This paper considers the evaluation of discretely distributed treatments when outcomes are only observed for a subpopulation due to sample selection or outcome attrition. For identification, we combine a selection-on-observables assumption for treatment assignment with either selection-on-observables or instrumental variable assumptions concerning the outcome attrition/sample selection process. We also consider dynamic confounding, meaning that covariates that jointly affect sample selection and the outcome may (at least partly) be influenced by the treatment. To control in a data-driven way for a potentially high dimensional set of pre- and/or post-treatment covariates, we adapt the double machine learning framework for treatment evaluation to sample selection problems. We make use of (a) Neyman-orthogonal, doubly robust, and efficient score functions, which imply the robustness of treatment effect estimation to moderate regularization biases in the machine learning-based estimation of the outcome, treatment, or sample selection models and (b) sample splitting (or cross-fitting) to prevent overfitting bias. We demonstrate that the proposed estimators are asymptotically normal and root-n consistent under specific regularity conditions concerning the machine learners and investigate their finite sample properties in a simulation study. We also apply our proposed methodology to the Job Corps data for evaluating the effect of training on hourly wages which are only observed conditional on employment. The estimator is available in the causalweight package for the statistical software R.

In the absence of effective vaccination, mass testing and quarantining of positive cases and their contacts could help to mitigate pandemics and allow economies to stay open. We investigate the effects of repeated mass testing on the COVID-19 pandemic caused by the SARS-CoV-2 virus, using data from the first ever nationwide rapid antigen testing implemented in Slovakia in autumn 2020. After the first round of testing, only districts above an ex ante unknown threshold of test positivity were re-tested. Comparing districts above and below the threshold, we provide evidence that repeated mass antigen testing can temporarily reduce the number of new infections. Our results suggest that mass testing coupled with the quarantining of positive cases and their contacts could be an effective tool in mitigating pandemics. For lasting effects, re-testing at regular intervals would likely be necessary.