Efficacy and Safety of Xanomeline-Trospium Chloride in Schizophrenia: A Randomized Clinical Trial.

Importance: A significant need exists for new antipsychotic medications with different mechanisms of action, greater efficacy, and better tolerability than existing agents. Xanomeline is a dual M1/M4 preferring muscarinic receptor agonist with no direct D2 dopamine receptor blocking activity. KarXT combines xanomeline with the peripheral muscarinic receptor antagonist trospium chloride with the goal of reducing adverse events due to xanomeline-related peripheral muscarinic receptor activation. In prior trials, xanomeline-trospium chloride was effective in reducing symptoms of psychosis and generally well tolerated in people with schizophrenia. Objective: To evaluate the efficacy and safety of xanomeline-trospium vs placebo in adults with schizophrenia. Design, Setting, and Participants: EMERGENT-3 (NCT04738123) was a phase 3, multicenter, randomized, double-blind, placebo-controlled, 5-week trial of xanomeline-trospium in people with schizophrenia experiencing acute psychosis, conducted between April 1, 2021, and December 7, 2022, at 30 inpatient sites in the US and Ukraine. Data were analyzed from February to June 2023. Interventions: Participants were randomized 1:1 to receive xanomeline-trospium chloride (maximum dose xanomeline 125 mg/trospium 30 mg) or placebo for 5 weeks. Main Outcomes and Measures: The prespecified primary end point was change from baseline to week 5 in Positive and Negative Syndrome Scale (PANSS) total score. Secondary outcome measures were change from baseline to week 5 in PANSS positive subscale score, PANSS negative subscale score, PANSS Marder negative factor score, Clinical Global Impression-Severity score, and proportion of participants with at least a 30% reduction in PANSS total score. Safety and tolerability were also evaluated. Results: A total of 256 participants (mean [SD] age, 43.1 [11.8] years; 191 men [74.6%]; 156 of 256 participants [60.9%] were Black or African American, 98 [38.3%] were White, and 1 [0.4%] was Asian) were randomized (125 in xanomeline-trospium group and 131 in placebo group). At week 5, xanomeline-trospium significantly reduced PANSS total score compared with placebo (xanomeline-trospium , -20.6; placebo, -12.2; least squares mean difference, -8.4; 95% CI, -12.4 to -4.3; P < .001; Cohen d effect size, 0.60). Discontinuation rates due to treatment-emergent adverse events (TEAEs) were similar between the xanomeline-trospium (8 participants [6.4%]) and placebo (7 participants [5.5%]) groups. The most common TEAEs in the xanomeline-trospium vs placebo group were nausea (24 participants [19.2%] vs 2 participants [1.6%]), dyspepsia (20 participants [16.0%] vs 2 participants [1.6%]), vomiting (20 participants [16.0%] vs 1 participant [0.8%]), and constipation (16 participants [12.8%] vs 5 participants [3.9%]). Measures of extrapyramidal symptoms, weight gain, and somnolence were similar between treatment groups. Conclusions and Relevance: Xanomeline-trospium was efficacious and well tolerated in people with schizophrenia experiencing acute psychosis. These findings, together with the previously reported and consistent results from the EMERGENT-1 and EMERGENT-2 trials, support the potential of xanomeline-trospium to be the first in a putative new class of antipsychotic medications without D2 dopamine receptor blocking activity. Trial Registration: ClinicalTrials.gov Identifier: NCT04738123.

Efficacy and safety of the muscarinic receptor agonist KarXT (xanomeline-trospium) in schizophrenia (EMERGENT-2) in the USA: results from a randomised, double-blind, placebo-controlled, flexible-dose phase 3 trial.

BACKGROUND: New treatments with new mechanisms are urgently needed for people with schizophrenia. Xanomeline is a dual M1 and M4-preferring muscarinic receptor agonist that does not block D2 dopamine receptors, unlike all currently approved treatments for schizophrenia. Xanomeline-trospium (KarXT) combines xanomeline with the peripherally restricted muscarinic receptor antagonist trospium chloride with the goal of ameliorating xanomeline-related adverse events associated with peripheral muscarinic receptors. The EMERGENT-2 trial aimed to assess the efficacy and safety of KarXT in people with schizophrenia experiencing acute psychosis. METHODS: EMERGENT-2 was a randomised, double-blind, placebo-controlled, flexible-dose, 5-week, inpatient, phase 3 trial in people with schizophrenia. Participants were adults aged 18-65 years with a diagnosis of schizophrenia who had a recent worsening of psychosis warranting hospital admission, a Positive and Negative Syndrome Scale (PANSS) score of 80 or higher, and a Clinical Global Impression-Severity score of 4 or higher. The participants were recruited from 22 inpatient sites in the USA, and were randomly assigned (1:1) to KarXT or placebo twice per day. Participants randomly assigned to KarXT received 50 mg xanomeline and 20 mg trospium twice per day for the first 2 days and then 100 mg xanomeline and 20 mg trospium twice per day for days 3-7. Beginning on day 8, KarXT dosing was flexible with an optional increase to 125 mg xanomeline and 30 mg trospium twice per day and the option to return to 100 mg xanomeline and 20 mg trospium based on tolerability. The primary endpoint was change from baseline to week 5 in PANSS total score. Efficacy analyses used the modified intention-to-treat population (all randomly assigned participants who received at least one trial medication dose and had at least one post-baseline PANSS assessment). Least squares mean change from baseline, SE, and least squares mean difference between the KarXT and placebo groups at week 5, along with the 95% CI and two-sided p values were calculated for the primary and secondary continuous efficacy endpoints. Safety analyses included all participants receiving at least one trial medication dose and used descriptive statistics. This trial is registered with ClinicalTrials.gov (NCT04659161). FINDINGS: From Dec 16, 2020, to April 13, 2022, of 407 people who were screened, 252 participants meeting enrolment criteria were randomly assigned to the KarXT (n=126) or placebo (n=126). Baseline PANSS total scores were 98·3 (KarXT; n=126) and 97·9 (placebo; n=125). The trial met the primary endpoint with a mean change from baseline to week 5 in PANSS total score that favoured KarXT (-21·2 points, SE 1·7) versus placebo (-11·6 points, 1·6; least squares mean difference -9·6; 95% CI -13·9 to -5·2; p<0·0001, Cohen's d effect size=0·61). All secondary endpoints were also met, and favoured KarXT versus placebo (p<0·05). The most common adverse events with KarXT versus placebo were constipation (27 [21%] vs 13 [10%]), dyspepsia (24 [19%] vs 10 [8%]), headache (17 [14%] vs 15 [12%]), nausea (24 [19%] vs seven [6%]), vomiting (18 [14%] vs one [1%]), hypertension (12 [10%] vs one [1%]), dizziness (11 [9%] vs four [3%]), gastro-oesophageal reflux disease (eight [6%] vs zero [0%]), and diarrhoea (seven [6%] vs four [3%]). Treatment-emergent adverse event rates of extrapyramidal motor symptoms (KarXT, zero [0%] vs placebo, zero [0%]), akathisia (one [1%] vs one [1%]), weight gain (zero [0%] vs one [1%]), and somnolence (six [5%] vs five [4%]) were similar between the KarXT and placebo groups, as were adverse event-related discontinuation rates (nine [7%] vs seven [6%]). INTERPRETATION: In the EMERGENT-2 trial, KarXT was effective in reducing positive and negative symptoms and was generally well tolerated. These results support the potential for KarXT to represent a new class of effective and well tolerated antipsychotic medicines based on activating muscarinic receptors, not the D2 dopamine receptor-blocking mechanism of all current antipsychotic medications. Results from additional trials, including the identical EMERGENT-3 trial and the 52-week, open-label EMERGENT-4 and EMERGENT-5 trials, will provide additional information on the efficacy and safety of KarXT in people with schizophrenia. FUNDING: Karuna Therapeutics.

Modeling personalized heart rate response to exercise and environmental factors with wearables data.

Heart rate (HR) response to workout intensity reflects fitness and cardiorespiratory health. Physiological models have been developed to describe such heart rate dynamics and characterize cardiorespiratory fitness. However, these models have been limited to small studies in controlled lab environments and are challenging to apply to noisy-but ubiquitous-data from wearables. We propose a hybrid approach that combines a physiological model with flexible neural network components to learn a personalized, multidimensional representation of fitness. The physiological model describes the evolution of heart rate during exercise using ordinary differential equations (ODEs). ODE parameters are dynamically derived via a neural network connecting personalized representations to external environmental factors, from area topography to weather and instantaneous workout intensity. Our approach efficiently fits the hybrid model to a large set of 270,707 workouts collected from wearables of 7465 users from the Apple Heart and Movement Study. The resulting model produces fitness representations that accurately predict full HR response to exercise intensity in future workouts, with a per-workout median error of 6.1 BPM [4.4-8.8 IQR]. We further demonstrate that the learned representations correlate with traditional metrics of cardiorespiratory fitness, such as VO2 max (explained variance 0.81 ± 0.003). Lastly, we illustrate how our model is naturally interpretable and explicitly describes the effects of environmental factors such as temperature and humidity on heart rate, e.g., high temperatures can increase heart rate by 10%. Combining physiological ODEs with flexible neural networks can yield interpretable, robust, and expressive models for health applications.

Impact of spirituality on patient memories of intensive care unit stays: A nationwide cross-sectional study.

Background: Absent or delusional memories are experienced by many patients following an intensive care unit (ICU) stay. Up to 70% may have delusional or hallucinatory intrusive memories, which may persist long term. This study aims to investigate how spiritual health (SH) impacts ICU patients' memories and quality of communication (QoC) between patients and physicians (PP) or nurses (PN). Methods: This cross-sectional study was conducted across the country on ICU patients discharged from 45 medical centers in 31 provinces of Iran, to evaluate the direct and indirect effects of SH and ICU characteristics on patients' memory. Two valid and standard ICU memory tools (ICU-MT) and SH questionnaires were administered to patients 1 day post-ICU discharge used. Results: No significant direct effect of SH scores on ICU-MT items was observed. No significant correlation was observed between PP-QoC and PN-QoC variables and primary items of the ICU-MT. Female sex positively correlated with the development of delusional memories (odds ratio [OR]: 1.730, 95% confidence interval [CI]: 1.025-2.915, P < 0.05). Subjects admitted to the medical ICU were less likely to remember being in the ICU (OR: 0.398, 95% CI: 0.159-0.996, P < 0.05), and were less likely to report intrusive memories from their time in the hospital or events that led to their admission (OR: 0.19, 95% CI: 0.086-0.419, P < 0.001). Conclusions: The results of this study indicate that the spiritual health indirectly increased coping with intrusive memories, however, no direct effect was observed on ICU-MT items. The quality of communication between patients and physicians and nurses significantly mediated development of intrusive memories.

Linear versus Turbulent Airflow Tracheostomy Heat and Moisture Exchangers: A Crossover Study.

OBJECTIVE: This study aimed to evaluate the impact of two tracheostomy heat and moisture exchangers (HMEs), namely the Shikani Oxygen HME™ (S-O2 HME, ball type, turbulent airflow) and Mallinckrodt Tracheolife II DAR HME (M-O2 HME; flapper type, linear airflow) on tracheobronchial mucosal health, oxygenation, humidification, and patient preference. METHODS: A randomized cross-over study was conducted with HME-naïve long-term tracheostomy subjects at two academic medical centers. Bronchoscopy assessments of mucosal health were performed at baseline and day 5 of HME application, along with oxygen saturation (SpO2 ) and breathed air humidity at four oxygen flow rates (1, 2, 3, and 5 lpm). Patient preference was assessed on study conclusion. RESULTS: Both HMEs were associated with improved mucosal inflammation and decreased mucus production (p < 0.0002), with greater improvements in the S-O2 HME group (p < 0.007). Both HMEs improved humidity concentration at each oxygen flow rate (p < 0.0001), without significant differences between groups. SpO2 was greater for the S-O2 HME versus the M-O2 HME across all measured oxygen flow rates (p = 0.003). At low oxygen flow rates (1 or 2 lpm), the SpO2 in the S-O2 HME group was similar to that of the M-O2 HME at higher oxygen flow rates (3 or 5 lpm; p = 0.6). Ninety percent of subjects preferred the S-O2 HME. CONCLUSION: Tracheostomy HME uses correlated with improved indicators of tracheobronchial mucosal health, humidity, and oxygenation. The S-O2 HME outperformed the M-O2 HME with respect to tracheobronchial inflammation, SpO2 , and patient preference. Regular HME use by tracheostomy patients is recommended to optimize pulmonary health. Newer ball-type speaking valve technology additionally allows concomitant HME and speaking valve application. LEVEL OF EVIDENCE: 2 Laryngoscope, 133:3422-3428, 2023.

Evidence of trospium's ability to mitigate cholinergic adverse events related to xanomeline: phase 1 study results.

RATIONALE: The M1/M4 preferring muscarinic receptor agonist xanomeline demonstrated antipsychotic and procognitive effects in patients with Alzheimer's disease or schizophrenia in prior studies, but further clinical development was limited by cholinergic adverse events (AEs). KarXT combines xanomeline with the peripherally restricted muscarinic receptor antagonist trospium with the goal of improving tolerability and is in clinical development for schizophrenia and other neuropsychiatric disorders. OBJECTIVE: Test the hypothesis that trospium can mitigate cholinergic AEs associated with xanomeline. METHODS: Healthy volunteers enrolled in this phase 1 (NCT02831231), single-site, 9-day, double-blind comparison of xanomeline alone (n = 33) versus KarXT (n = 35). Rates of five prespecified cholinergic AEs (nausea, vomiting, diarrhea, excessive sweating, salivary hypersecretion) were compared between treatment arms. Vital signs, electrocardiograms (ECGs), safety laboratory values, and pharmacokinetic (PK) analyses were assessed. A self-administered visual analog scale (VAS) and clinician-administered scales were employed. RESULTS: Compared with xanomeline alone, KarXT reduced composite incidences of the five a priori selected cholinergic AEs by 46% and each individual AE by ≥ 29%. There were no episodes of syncope in KarXT-treated subjects; two cases occurred in the xanomeline-alone arm. The rate of postural dizziness was 11.4% in the KarXT arm versus 27.2% with xanomeline alone. ECG, vital signs, and laboratory values were not meaningfully different between treatment arms. The VAS and clinician-administered scales tended to favor KarXT. PK analysis revealed that trospium did not affect xanomeline's PK profile. CONCLUSIONS: Trospium was effective in mitigating xanomeline-related cholinergic AEs. KarXT had an improved safety profile compared with xanomeline alone.

Safety and tolerability of KarXT (xanomeline-trospium) in a phase 2, randomized, double-blind, placebo-controlled study in patients with schizophrenia.

KarXT combines xanomeline, an M1/M4 preferring muscarinic agonist with no direct D2 receptor antagonism, with the peripherally restricted anticholinergic trospium. In EMERGENT-1 (NCT03697252), a 5-week, randomized, double-blind, placebo-controlled, phase 2 study in inpatients with schizophrenia, KarXT met the primary efficacy endpoint, numerous secondary endpoints, and was generally well tolerated. Here, we conducted additional post hoc analyses of safety and tolerability data of KarXT from EMERGENT-1 with a particular focus on adverse events (AEs) that may be associated with muscarinic receptor agonism (nausea or vomiting) or antagonism (dry mouth or constipation). A total of 179 patients received at least one dose of either KarXT (n = 89) or placebo (n = 90) and were included in the analyses. KarXT was associated with a low overall AE burden. The majority of procholinergic and anticholinergic AEs with KarXT were mild, occurred in the first 1-2 weeks of treatment, and were transient with a median duration ranging from 1 day for vomiting to 13 days for dry mouth. No patients in either treatment group discontinued the study due to any procholinergic or anticholinergic AEs. Incidence of somnolence/sedation AEs with KarXT were low and similar to those in the placebo group. KarXT was associated with no significant or clinically relevant changes in body weight, metabolic parameters, or vital signs. KarXT was generally well tolerated with an AE profile consistent with the activity of xanomeline-trospium at muscarinic receptors.

Effectiveness of KarXT (xanomeline-trospium) for cognitive impairment in schizophrenia: post hoc analyses from a randomised, double-blind, placebo-controlled phase 2 study.

The muscarinic receptor agonist xanomeline improved cognition in phase 2 trials in Alzheimer's disease and schizophrenia. We present data on the effect of KarXT (xanomeline-trospium) on cognition in schizophrenia from the 5-week, randomised, double-blind, placebo-controlled EMERGENT-1 trial (NCT03697252). Analyses included 125 patients with computerised Cogstate Brief Battery (CBB) subtest scores at baseline and endpoint. A post hoc subgroup analysis evaluated the effects of KarXT on cognitive performance in patients with or without clinically meaningful cognitive impairment at baseline, and a separate outlier analysis excluded patients with excessive intraindividual variability (IIV) across cognitive subdomains. ANCOVA models assessed treatment effects for completers and impairment subgroups, with or without removal of outliers. Sample-wide, cognitive improvement was numerically but not statistically greater with KarXT (n = 60) than placebo (n = 65), p = 0.16. However, post hoc analyses showed 65 patients did not exhibit clinically meaningful cognitive impairment at baseline, while eight patients had implausibly high IIV at one or both timepoints. Significant treatment effects were observed after removing outliers (KarXT n = 54, placebo n = 63; p = 0.04). Despite the small sample size, a robust (d = 0.50) and significant effect was observed among patients with cognitive impairment (KarXT n = 23, placebo n = 37; p = 0.03). These effects did not appear to be related to improvement in PANSS total scores (linear regression, R2 = 0.03). Collectively, these findings suggest that KarXT may have a separable and meaningful impact on cognition, particularly among patients with cognitive impairment.

Muscarinic Acetylcholine Receptor Agonists as Novel Treatments for Schizophrenia.

Schizophrenia remains a challenging disease to treat effectively with current antipsychotic medications due to their limited efficacy across the entire spectrum of core symptoms as well as their often burdensome side-effect profiles and poor tolerability. An unmet need remains for novel, mechanistically unique, and better tolerated therapeutic agents for treating schizophrenia, especially those that treat not only positive symptoms but also the negative and cognitive symptoms of the disease. Almost 25 years ago, the muscarinic acetylcholine receptor (mAChR) agonist xanomeline was reported to reduce psychotic symptoms and improve cognition in patients with Alzheimer's disease. The antipsychotic and procognitive properties of xanomeline were subsequently confirmed in a small study of acutely psychotic patients with chronic schizophrenia. These unexpected clinical findings have prompted considerable efforts across academia and industry to target mAChRs as a new approach to potentially treat schizophrenia and other psychotic disorders. The authors discuss recent advances in mAChR biology and pharmacology and the current understanding of the relative roles of the various mAChR subtypes, their downstream cellular effectors, and key neural circuits mediating the reduction in the core symptoms of schizophrenia in patients treated with xanomeline. They also provide an update on the status of novel mAChR agonists currently in development for potential treatment of schizophrenia and other neuropsychiatric disorders.

Antipsychotic Efficacy of KarXT (Xanomeline-Trospium): Post Hoc Analysis of Positive and Negative Syndrome Scale Categorical Response Rates, Time Course of Response, and Symptom Domains of Response in a Phase 2 Study.

Objective: To evaluate Positive and Negative Syndrome Scale (PANSS) categorical response rates, time course of response, and symptom subdomains of response with the combination oral agent KarXT (xanomeline-trospium) in the treatment of schizophrenia.Methods: Post hoc analysis was conducted for EMERGENT-1 (NCT03697252), a 5-week, inpatient, placebo-controlled, phase 2 study of acute psychosis in patients who met DSM-5 criteria for schizophrenia. The EMERGENT-1 study was conducted between September 2018 and August 2019. Categorical thresholds of response used were PANSS total score reductions of ≥ 20%, ≥ 30%, ≥ 40%, and ≥ 50% between baseline and study end. Number needed to treat (NNT) for each categorical threshold was calculated. The proportion of KarXT- and placebo-treated patients achieving each response threshold at weeks 2, 4, and 5 was assessed. Marder 5-factor analysis of PANSS assessed response with KarXT across symptom domains.Results: A total of 83 patients in the KarXT group and 87 patients in the placebo group were included in the modified intent-to-treat analysis. Response rates with KarXT ranged from 59.0% for a ≥ 20% threshold to 15.7% for a ≥ 50% threshold. All response rates with KarXT were significantly higher than in the placebo arm (P < .05), with NNTs ranging from 3 (≥ 20% improvement) to 11 (≥ 50% improvement). KarXT was associated with a significantly higher response rate relative to placebo as early as 2 weeks for ≥ 20% (P = .0001) and ≥ 30% (P = .0022) thresholds and at 4 weeks for the ≥ 40% (P = .0049) and ≥ 50% (P = .0041) thresholds. Each of the Marder 5 factors showed significant differences favoring KarXT over placebo (P < .05) by 2 weeks and continuing through week 5 (endpoint Cohen d effect sizes, 0.48-0.66).Conclusions: KarXT provided clinically meaningful responder rates on PANSS total score compared with placebo at each response threshold, providing further support of the successful primary and secondary endpoints. Response was demonstrated as early as 2 weeks relative to placebo. KarXT demonstrated improvements vs placebo in all 5 factors (positive symptoms, negative symptoms, disorganized thought, uncontrolled hostility, and anxiety/depression).Trial Registration: ClinicalTrials.gov identifier: NCT03697252.

Statistical Deconvolution for Inference of Infection Time Series.

Accurate measurement of daily infection incidence is crucial to epidemic response. However, delays in symptom onset, testing, and reporting obscure the dynamics of transmission, necessitating methods to remove the effects of stochastic delays from observed data. Existing estimators can be sensitive to model misspecification and censored observations; many analysts have instead used methods that exhibit strong bias. We develop an estimator with a regularization scheme to cope with stochastic delays, which we term the robust incidence deconvolution estimator. We compare the method to existing estimators in a simulation study, measuring accuracy in a variety of experimental conditions. We then use the method to study COVID-19 records in the United States, highlighting its stability in the face of misspecification and right censoring. To implement the robust incidence deconvolution estimator, we release incidental, a ready-to-use R implementation of our estimator that can aid ongoing efforts to monitor the COVID-19 pandemic.

Risk factors, time to onset and recurrence of delirium in a mixed medical-surgical ICU population: A secondary analysis using Cox and CHAID decision tree modeling.

A retrospective secondary analysis of 4,200 patients was collected from two academic medical centers. Delirium was assessed using the Confusion Assessment Method for the Intensive Care Unit (CAM-ICU) in all patients. Univariate and multivariate Cox models, logistic regression analysis, and Chi-square Automatic Interaction Detector (CHAID) decision tree modeling were used to explore delirium risk factors. Increased delirium risk was associated with exposed only to artificial light (AL) hazard ratio (HR) 1.84 (95 % CI: 1.66-2.044, P<0.001), physical restraint application 1.11 (95 % CI: 1.001-1.226, P=0.049), and high nursing care requirements (>8 hours per 8-hour shift) 1.18 (95 % CI: 1.048-1.338, P=0.007). Delirium incidence was inversely associated with greater family engagement 0.092 (95 % CI: 0.014-0.596, P=0.012), low staff burnout and anticipated turnover scores 0.093 (95 % CI: 0.014-0.600, P=0.013), non-ICU length-of-stay (LOS)<15 days 0.725 (95 % CI: 0.655-0.804, P<0.001), and ICU LOS ≤15 days 0.509 (95 % CI: 0.456-0.567, P<0.001). CHAID modeling indicated that AL exposure and age <65 years were associated with a high risk of delirium incidence, whereas SOFA score ≤11, APACHE IV score >15 and natural light (NL) exposure were associated with moderate risk, and female sex was associated with low risk. More rapid time to delirium onset correlated with baseline sleep disturbance (P=0.049), high nursing care requirements (P=0.019), and prolonged ICU and non-ICU hospital LOS (P<0.001). Delirium recurrence correlated with age >65 years (HR 2.198; 95 % CI: 1.101-4.388, P=0.026) and high nursing care requirements (HR 1.978, 95 % CI: 1.096-3.569), with CHAID modeling identifying AL exposure (P<0.001) and age >65 years (P=0.032) as predictive variables. Development of ICU delirium correlated with application of physical restraints, high nursing care requirements, prolonged ICU and non-ICU LOS, exposure exclusively to AL (rather than natural), less family engagement, and greater staff burnout and anticipated turnover scores. ICU delirium occurred more rapidly in patients with baseline sleep disturbance, and recurrence correlated with the presence of delirium on ICU admission, exclusive AL exposure, and high nursing care requirements.

Learning Invariant Representations with Missing Data.

Spurious correlations allow flexible models to predict well during training but poorly on related test populations. Recent work has shown that models that satisfy particular independencies involving correlation-inducing nuisance variables have guarantees on their test performance. Enforcing such independencies requires nuisances to be observed during training. However, nuisances, such as demographics or image background labels, are often missing. Enforcing independence on just the observed data does not imply independence on the entire population. Here we derive MMD estimators used for invariance objectives under missing nuisances. On simulations and clinical data, optimizing through these estimates achieves test performance similar to using estimators that make use of the full data.