Resistance-only and concurrent exercise induce similar myofibrillar protein synthesis rates and associated molecular responses in moderately active men before and after training.

Aerobic and resistance exercise (RE) induce distinct molecular responses. One hypothesis is that these responses are antagonistic and unfavorable for the anabolic response to RE when concurrent exercise is performed. This thesis may also depend on the participants' training status and concurrent exercise order. We measured free-living myofibrillar protein synthesis (MyoPS) rates and associated molecular responses to resistance-only and concurrent exercise (with different exercise orders), before and after training. Moderately active men completed one of three exercise interventions (matched for age, baseline strength, body composition, and aerobic capacity): resistance-only exercise (RE, n = 8), RE plus high-intensity interval exercise (RE+HIIE, n = 8), or HIIE+RE (n = 9). Participants trained 3 days/week for 10 weeks; concurrent sessions were separated by 3 h. On the first day of Weeks 1 and 10, muscle was sampled immediately before and after, and 3 h after each exercise mode and analyzed for molecular markers of MyoPS and muscle glycogen. Additional muscle, sampled pre- and post-training, was used to determine MyoPS using orally administered deuterium oxide (D2 O). In both weeks, MyoPS rates were comparable between groups. Post-exercise changes in proteins reflective of protein synthesis were also similar between groups, though MuRF1 and MAFbx mRNA exhibited some exercise order-dependent responses. In Week 10, exercise-induced changes in MyoPS and some genes (PGC-1ɑ and MuRF1) were dampened from Week 1. Concurrent exercise (in either order) did not compromise the anabolic response to resistance-only exercise, before or after training. MyoPS rates and some molecular responses to exercise are diminished after training.

The effect of sleep restriction, with or without high-intensity interval exercise, on behavioural alertness and mood state in young healthy males.

Mood state and alertness are negatively affected by sleep loss, and can be positively influenced by exercise. However, the potential mitigating effects of exercise on sleep-loss-induced changes in mood state and alertness have not been studied comprehensively. Twenty-four healthy young males were matched into one of three, 5-night sleep interventions: normal sleep (NS; total sleep time (TST) per night = 449 ± 22 min), sleep restriction (SR; TST = 230 ± 5 min), or sleep restriction and exercise (SR + EX; TST = 235 ± 5 min, plus three sessions of high-intensity interval exercise (HIIE)). Mood state was assessed using the profile of mood states (POMS) and a daily well-being questionnaire. Alertness was assessed using psychomotor vigilance testing (PVT). Following the intervention, POMS total mood disturbance scores significantly increased for both the SR and SR + EX groups, and were greater than the NS group (SR vs NS; 31.0 ± 10.7 A.U., [4.4-57.7 A.U.], p = 0.020; SR + EX vs NS; 38.6 ± 14.9 A.U., [11.1-66.1 A.U.], p = 0.004). The PVT reaction times increased in the SR (p = 0.049) and SR + EX groups (p = 0.033) and the daily well-being questionnaire revealed increased levels of fatigue in both groups (SR; p = 0.041, SR + EX; p = 0.026) during the intervention. Despite previously demonstrated physiological benefits of performing three sessions of HIIE during five nights of sleep restriction, the detriments to mood, wellness, and alertness were not mitigated by exercise in this study. Whether alternatively timed exercise sessions or other exercise protocols could promote more positive outcomes on these factors during sleep restriction requires further research.

Comment on Oberman & Vink: Should we fix or simulate the complete data in simulation studies evaluating missing data methods?

For simulation studies that evaluate methods of handling missing data, we argue that generating partially observed data by fixing the complete data and repeatedly simulating the missingness indicators is a superficially attractive idea but only rarely appropriate to use.

Longitudinal Internal Training Load and Exposure in a High-Performance Basketball Academy.

ABSTRACT: Lever, JR, Duffield, R, Murray, A, Bartlett, JD, and Fullagar, HHK. Longitudinal internal training load and exposure in a high-performance basketball academy. J Strength Cond Res 38(8): 1464-1471, 2024-This study describes the longitudinal training exposure (session counts) and internal training load (Rating of Perceived Exertion [RPE] and Session Rating of Perceived Exertion [sRPE]) of youth basketball players at a high-performance academy, based on the training year, training term, and playing position. Historical internal training load and training exposure data were collated from 45 male high-performance youth basketball athletes between 2015 and 2019. Data included session duration, RPE, sRPE, training type, and date. Linear mixed models and pairwise comparisons were performed on the weekly means and categorized by training year (year 1, year 2, year 3), term (term 1, term 2, term 3, term 4), and playing position (Backcourt, Frontcourt). Linear mixed models indicate that the individual athlete had the greatest influence on variance in training load and exposure. Significant differences were observed for increased session count, duration, and sRPE ( p < 0.001) in year 2 compared with year 1. These measures also increased within each year whereby term 3 and term 4 ( p < 0.001) were significantly greater than term 1 and term 2. No significant differences were observed between playing position ( p > 0.05). Training exposure and internal training load increase in year 2 from year 1 for high-performance youth basketball academy athletes. Differences between training load and exposure for terms (i.e., training blocks) suggest the phase of season influences training prescription, while playing position has limited effect.

An updated perspective on teleonomy.

The analysis of proxy failure given by John et al. provides a good starting point for interdisciplinary discussions. Here, the discussion of teleonomy is extended and updated to include more recent discourse on the topic.

Random with Respect to Fitness or External Selection? An Important but Often Overlooked Distinction.

Mutations are often described as being "random with respect to fitness." Here we show that the experiments used to establish randomness with respect to fitness are only capable of showing that mutations are random with respect to current external selection. Current debates about whether or not mutations are directed may be at least partially resolved by making use of this distinction. Additionally, this distinction has important mathematical, experimental, and inferential implications.

CT-guided vs. fluoroscopically guided transforaminal epidural steroid injections for lumbar radiculopathy: a comparison of efficacy, safety and cost.

INTRODUCTION: There are no formal guidelines for whether CT-guided or fluoroscopy-guided TFESI should be undertaken for patients with symptoms of lumbar nerve root irritation and corresponding nerve impingement. Here, we sought to compare the efficacy, safety and cost of computer tomography (CT)-guided and fluoroscopically guided transforaminal epidural steroid injection (TFESI). MATERIALS AND METHODS: All patients who underwent lumbar TFESI at our institution between June 2016 and June 2018 were identified. Six-week follow-up outcomes were categorised. The radiation doses and associated cost was retrieved from our institution's costing system. RESULTS: One hundred and sixteen patients were included (CT-50; fluoroscopy-56). There were no complications. More patients were discharged 6 weeks after CT-guided lumbar TFESI when compared with fluoroscopically guided TFESI (CT-23, fluoroscopy-14 (P = 0.027)). There was no difference in the number of patients who were referred to surgery (P = 0.18), for further pain management (P = 0.45), or for further TFESI (P = 0.43). The effective radiation dose was significantly higher for CT-guided TFESI (CT-5.73 mSv (3.87 to 7.76); fluoroscopy-0.55 mSv (0.11 to 1.4) (P < 0.01)). The total cost for CT-guided lumbar TFESI was £237.50 (£235 to £337), over £800 less than under fluoroscopic guidance (£1052 (£892.80 to £1298.00), P < 0.01)). Removing cost associated with staff and theatre use (staffing, theatre, medical indemnity and overheads) revealed CT-guided lumbar TFESI to be less expensive than if the procedure was fluoroscopy-guided-CT-guided: £132.6 (130.8 to 197.5); fluoroscopy: £237.4 (£209.2 to £271.9) (P = 0.019). CONCLUSIONS: CT-guided TFESI was associated with a higher discharge rate, a lower cost, but a ten times higher radiation dose when compared with fluoroscopically guided TFESI. Prospective studies are required to compare the efficacy of these procedures and to investigate how the radiation dose of CT-guided TFESI can be reduced without jeopardising efficacy or safety.

Standard and reference-based conditional mean imputation.

Clinical trials with longitudinal outcomes typically include missing data due to missed assessments or structural missingness of outcomes after intercurrent events handled with a hypothetical strategy. Approaches based on Bayesian random multiple imputation and Rubin's rules for pooling results across multiple imputed data sets are increasingly used in order to align the analysis of these trials with the targeted estimand. We propose and justify deterministic conditional mean imputation combined with the jackknife for inference as an alternative approach. The method is applicable to imputations under a missing-at-random assumption as well as for reference-based imputation approaches. In an application and a simulation study, we demonstrate that it provides consistent treatment effect estimates with the Bayesian approach and reliable frequentist inference with accurate standard error estimation and type I error control. A further advantage of the method is that it does not rely on random sampling and is therefore replicable and unaffected by Monte Carlo error.

Effect of Divergent Solar Radiation Exposure With Outdoor Versus Indoor Training in the Heat: Implications for Performance.

ABSTRACT: O'Connor, FK, Doering, TM, Minett, GM, Reaburn, PR, Bartlett, JD and Coffey, VG. Effect of divergent solar radiation exposure with outdoor versus indoor training in the heat: implications for performance. J Strength Cond Res 36(6): 1622-1628, 2022-The aim of this study was to determine physiological and perceptual responses and performance outcomes when completing high-intensity exercise in outdoor and indoor hot environments with contrasting solar radiation exposure. Seven cyclists and 9 Australian Football League (AFL) players undertook cycling trials in hot conditions (≥30 °C) outdoors and indoors. Cyclists completed 5 × 4 minutes intervals (∼80% peak power output [PPO]) with 2 minutes recovery (∼40% PPO) before a 20-km self-paced ride. Australian Football League players completed a standardized 20 minutes warm-up (∼65% mean 4-minute power output) then 5 × 2 minutes maximal effort intervals. Heart rate (HR), PO, ratings of perceived exertion (RPE), thermal comfort (TC), and thermal sensation (TS) were recorded. Core (Tc) and skin temperature (Tsk) were monitored in cyclists alone. In both studies, ambient temperature, relative humidity, and solar radiation were monitored outdoors and matched for ambient temperature and relative humidity indoors, generating different wet bulb globe temperature (WBGT) for cyclists, but the similar WBGT for AFL players through higher relative humidity indoors. The statistical significance was set at p ≤ 0.05. Cyclists' HR (p = 0.05), Tc (p = 0.03), and Tsk (p = 0.03) were higher outdoors with variable effects for increased RPE, TS, and TC (d = 0.2-1.3). Power output during intervals was not different between trials, but there were small-moderate improvements in cyclists' PO and 20-km time indoors (d = 0.3-0.6). There was a small effect (d = 0.2) for AFL players' mean PO to increase outdoors for interval 4 alone (p = 0.04); however, overall there were small-moderate effects for lower RPE and TS indoors (d = 0.2-0.5). Indoor training in hot conditions without solar radiation may promote modest reductions in physiological strain and improve performance capacity in well-trained athletes.

Multiple iliopsoas tendons: a cadaveric study and treatment implications for internal snapping hip syndrome.

PURPOSE: This cadaveric study aimed at describing the anatomical variations of the iliopsoas complex. METHODS: The iliopsoas complex was dissected unilaterally in 28 formalin-embalmed cadavers-13 males and 15 females with a mean age of 85.6 years. The number, courses and widths of the iliacus and psoas major tendons were determined. Patients with previous hip surgery were excluded. The following measurements were taken from the mid-inguinal point: the distance to the point of union of the psoas major and iliacus tendon; and the distance to the most distal insertion of iliopsoas. RESULTS: The presence of single, double and triple tendon insertions of iliopsoas were found in 12, 12 and 4 of the 28 specimens, respectively. When present, double and triple tendons inserted separately onto the lesser trochanter. The average length of the iliopsoas tendon from the mid-inguinal point to the most distal attachment at the lesser trochanter was 122.3 ± 13.0 mm. The iliacus muscle bulk merged with psoas major at an average distance of 24.9 ± 17.9 mm proximal to the mid-inguinal point. In all cases, the lateral-most fibres of iliacus yielded a non-tendinous, muscular insertion on to the anterior surface of the lesser trochanter and the femoral shaft, rather than joining onto the main iliopsoas tendon(s). The average total width of the psoas major tendon decreased with an increasing number of tendons: 14.6 ± 2.2 mm (single tendon), 8.2 ± 3.0 mm (2 tendons present) and 5.9 ± 1.1 mm (3 tendons present) (P < 0.001). CONCLUSIONS: The results of this study suggest that multiple tendinous insertions of iliopsoas are present as an anatomical variant in more than 50% of the population. The non-tendinous muscular insertion of the iliopsoas on to the anterior surface of the lesser trochanter and femoral shaft found represents a novel anatomical variant not previously described. LEVEL OF EVIDENCE: Level V.

The effect of sleep restriction, with or without high-intensity interval exercise, on myofibrillar protein synthesis in healthy young men.

KEY POINTS: Sleep restriction has previously been associated with the loss of muscle mass in both human and animal models. The rate of myofibrillar protein synthesis (MyoPS) is a key variable in regulating skeletal muscle mass and can be increased by performing high-intensity interval exercise (HIIE), although the effect of sleep restriction on MyoPS is unknown. In the present study, we demonstrate that participants undergoing a sleep restriction protocol (five nights, with 4 h in bed each night) had lower rates of skeletal muscle MyoPS; however, rates of MyoPS were maintained at control levels by performing HIIE during this period. Our data suggest that the lower rates of MyoPS in the sleep restriction group may contribute to the detrimental effects of sleep loss on muscle mass and that HIIE may be used as an intervention to counteract these effects. ABSTRACT: The present study aimed to investigate the effect of sleep restriction, with or without high-intensity interval exercise (HIIE), on the potential mechanisms underpinning previously-reported sleep-loss-induced reductions to muscle mass. Twenty-four healthy, young men underwent a protocol consisting of two nights of controlled baseline sleep and a five-night intervention period. Participants were allocated into one of three parallel groups, matched for age, V̇O2peak , body mass index and habitual sleep duration; a normal sleep (NS) group [8 h time in bed (TIB) each night], a sleep restriction (SR) group (4 h TIB each night), and a sleep restriction and exercise group (SR+EX, 4 h TIB each night, with three sessions of HIIE). Deuterium oxide was ingested prior to commencing the study and muscle biopsies obtained pre- and post-intervention were used to assess myofibrillar protein synthesis (MyoPS) and molecular markers of protein synthesis and degradation signalling pathways. MyoPS was lower in the SR group [fractional synthetic rate (% day-1 ), mean ± SD, 1.24 ± 0.21] compared to both the NS (1.53 ± 0.09) and SR+EX groups (1.61 ± 0.14) (P < 0.05). However, there were no changes in the purported regulators of protein synthesis (i.e. p-AKTser473 and p-mTORser2448 ) and degradation (i.e. Foxo1/3 mRNA and LC3 protein) in any group. These data suggest that MyoPS is acutely reduced by sleep restriction, although MyoPS can be maintained by performing HIIE. These findings may explain the sleep-loss-induced reductions in muscle mass previously reported and also highlight the potential therapeutic benefit of HIIE to maintain myofibrillar remodelling in this context.

Robustness of ANCOVA in randomized trials with unequal randomization.

Randomized trials with continuous outcomes are often analyzed using analysis of covariance (ANCOVA), with adjustment for prognostic baseline covariates. The ANCOVA estimator of the treatment effect is consistent under arbitrary model misspecification. In an article recently published in the journal, Wang et al proved the model-based variance estimator for the treatment effect is also consistent under outcome model misspecification, assuming the probability of randomization to each treatment is 1/2. In this reader reaction, we derive explicit expressions which show that when randomization is unequal, the model-based variance estimator can be biased upwards or downwards. In contrast, robust sandwich variance estimators can provide asymptotically valid inferences under arbitrary misspecification, even when randomization probabilities are not equal.

Structure of malaria invasion protein RH5 with erythrocyte basigin and blocking antibodies.

Invasion of host erythrocytes is essential to the life cycle of Plasmodium parasites and development of the pathology of malaria. The stages of erythrocyte invasion, including initial contact, apical reorientation, junction formation, and active invagination, are directed by coordinated release of specialized apical organelles and their parasite protein contents. Among these proteins, and central to invasion by all species, are two parasite protein families, the reticulocyte-binding protein homologue (RH) and erythrocyte-binding like proteins, which mediate host-parasite interactions. RH5 from Plasmodium falciparum (PfRH5) is the only member of either family demonstrated to be necessary for erythrocyte invasion in all tested strains, through its interaction with the erythrocyte surface protein basigin (also known as CD147 and EMMPRIN). Antibodies targeting PfRH5 or basigin efficiently block parasite invasion in vitro, making PfRH5 an excellent vaccine candidate. Here we present crystal structures of PfRH5 in complex with basigin and two distinct inhibitory antibodies. PfRH5 adopts a novel fold in which two three-helical bundles come together in a kite-like architecture, presenting binding sites for basigin and inhibitory antibodies at one tip. This provides the first structural insight into erythrocyte binding by the Plasmodium RH protein family and identifies novel inhibitory epitopes to guide design of a new generation of vaccines against the blood-stage parasite.

The learning curves of a validated virtual reality hip arthroscopy simulator.

INTRODUCTION: Decreases in trainees' working hours, coupled with evidence of worse outcomes when hip arthroscopies are performed by inexperienced surgeons, mandate an additional means of training. Though virtual reality simulation has been adopted by other surgical specialities, its slow uptake in arthroscopic training is due to a lack of evidence as to its benefits. These benefits can be demonstrated through learning curves associated with simulator training-with practice reflecting increases in validated performance metrics. METHODS: Twenty-five medical students with no previous experience of hip arthroscopy completed seven weekly simulated arthroscopies of a healthy virtual hip joint using a 70° arthroscope in the supine position. Twelve targets were visualised within the central compartment, six via the anterior portal, three via the anterolateral portal and three via the posterolateral portal. Task duration, number of collisions (bone and soft-tissue), and distance travelled by arthroscope were measured by the simulator for every session of each student. RESULTS: Learning curves were demonstrated by the students, with improvements in time taken, number of collisions (bone and soft-tissue), collision length and efficiency of movement (all p < 0.01). Improvements in time taken, efficiency of movement and number of collisions with soft-tissue were first seen in session 3 and improvements in all other parameters were seen in session 4. No differences were found after session 5 for time taken and length of soft-tissue collision. No differences in number of collisions (bone and soft-tissue), length of collisions with bone, and efficiency of movement were found after session 6. CONCLUSIONS: The results of this study demonstrate learning curves for a hip arthroscopy simulator, with significant improvements seen after three sessions. All performance metrics were found to improved, demonstrating sufficient visuo-haptic consistency within the virtual environment, enabling individuals to develop basic arthroscopic skills.

Virtual reality hip arthroscopy simulator demonstrates sufficient face validity.

PURPOSE: To test the face validity of the hip diagnostics module of a virtual reality hip arthroscopy simulator. METHODS: A total of 25 orthopaedic surgeons, 7 faculty members and 18 orthopaedic residents, performed diagnostic supine hip arthroscopies of a healthy virtual reality hip joint using a 70° arthroscope. Twelve specific targets were visualised within the central compartment; six via the anterior portal, three via the anterolateral portal and three via the posterolateral portal. This task was immediately followed by a questionnaire regarding the realism and training capability of the system. This consisted of seven questions addressing the verisimilitude of the simulator and five questions addressing the training environment of the simulator. Each question consisted of a statement stem and 10-point Likert scale. Following similar work in surgical simulators, a rating of 7 or above was considered an acceptable level of realism. RESULTS: The diagnostic hip arthroscopy module was found to have an acceptable level of realism in all domains apart from the tactile feedback received from the soft tissue. 23 out of 25 participants (92%) felt the simulator provided a non-threatening learning environment and 22 participants (88%) stated they enjoyed using the simulator. It was most frequently agreed that the level of trainees who would benefit most from the simulator were registrars and fellows (22 participants; 88%). Additionally, 21 of the participants (84%) agreed that this would be a beneficial training modality for foundation and core trainees, and 20 participants (80%) agreed that his would be beneficial for consultants. CONCLUSIONS: This VR hip arthroscopy simulator was demonstrated to have a sufficient level of realism, thus establishing its face validity. These results suggest this simulator has sufficient realism for use in the acquisition of basic arthroscopic skills and supports its use in orthopaedics surgical training. LEVEL OF EVIDENCE: I.

Multiple imputation of missing data in nested case-control and case-cohort studies.

The nested case-control and case-cohort designs are two main approaches for carrying out a substudy within a prospective cohort. This article adapts multiple imputation (MI) methods for handling missing covariates in full-cohort studies for nested case-control and case-cohort studies. We consider data missing by design and data missing by chance. MI analyses that make use of full-cohort data and MI analyses based on substudy data only are described, alongside an intermediate approach in which the imputation uses full-cohort data but the analysis uses only the substudy. We describe adaptations to two imputation methods: the approximate method (MI-approx) of White and Royston (2009) and the "substantive model compatible" (MI-SMC) method of Bartlett et al. (2015). We also apply the "MI matched set" approach of Seaman and Keogh (2015) to nested case-control studies, which does not require any full-cohort information. The methods are investigated using simulation studies and all perform well when their assumptions hold. Substantial gains in efficiency can be made by imputing data missing by design using the full-cohort approach or by imputing data missing by chance in analyses using the substudy only. The intermediate approach brings greater gains in efficiency relative to the substudy approach and is more robust to imputation model misspecification than the full-cohort approach. The methods are illustrated using the ARIC Study cohort. Supplementary Materials provide R and Stata code.

What Is the Risk Posed to the Lateral Femoral Cutaneous Nerve During the Use of the Anterior Portal of Supine Hip Arthroscopy and the Minimally Invasive Anterior Approach for Total Hip Arthroplasty?

PURPOSE: To determine: (1) What is the proximity of the lateral femoral cutaneous nerve (LFCN) to the anterior portal (AP) used in supine hip arthroscopy? (2) What is the proximity of the LCFN to the incision in the minimally invasive anterior approach (MIAA) for total hip arthroplasty? (3) What effect does lateralizing the AP have on the likelihood of nerve injury? (4) What branching patterns are observable in the LFCN? METHODS: Forty-five hemipelves were dissected. The LFCN was identified and its path dissected. The positions of the nerve in relation to the AP and the MIAA incision were measured. RESULTS: The AP intersected with 38% of nerves. In the remainder, the LFCN was located 5.7 ± 4.5 mm from the portal's edge. In addition, 44% of nerves crossed the incision of the MIAA. Of those that did not, the average minimum distance from the incision was 14.4 ± 7.0 mm. We found a significant reduction in risk if the AP is moved medially by 5 mm or laterally by 15 mm (P = .0054 and P = .0002). The LFCN showed considerable variation with 4 branching variants. CONCLUSIONS: These results show that the LFCN is at high risk during supine hip arthroscopy and the MIAA, emphasizing the need for meticulous dissection. We suggest that relocation of the AP 5 mm medially or 15 mm laterally will reduce the risk to the LFCN. CLINICAL RELEVANCE: These findings should aid surgeons in minimizing the risk to the LCFN during hip arthroscopy and the minimally invasive anterior approach to the hip.

Differential training loads and individual fitness responses to pre-season in professional rugby union players.

We aimed to compare differentiated training loads (TL) between fitness responders and non-responders to an eight-week pre-season training period in a squad of thirty-five professional rugby union players. Differential TL were calculated by multiplying player's perceptions of breathlessness (sRPE-B) and leg muscle exertion (sRPE-L) with training duration for each completed session. Performance-based fitness measures included the Yo-Yo Intermittent Recovery Test Level 1 (YYIRTL1), 10-, 20-, and 30-m linear sprint times, countermovement jump height (CMJ) and predicted one-repetition maximum back squat (P1RM Squat). The proportion of responders (≥ 75% chance that the observed change in fitness was > typical error and smallest worthwhile change) were 37%, 50%, 52%, 82% and 70% for YYIRTL1, 20/30-m, 10-m, CMJ and P1RM Squat, respectively. Weekly sRPE-B-TL was very likely higher in YYIRTL1 responders (mean difference = 18%; ±90% confidence limits 11%), likely lower in 20/30-m (19%; ±20%) and 10-m (18%; ±17%) responders, and likely higher in CMJ responders (15%; ±16%). All other comparisons were unclear. Weekly sRPE-B discriminate between rugby union players who respond to pre-season training when compared with players who do not. Our findings support the collection of differential ratings of perceived exertion and the use of individual response analysis in team-sport athletes.

Covariate adjustment and estimation of mean response in randomised trials.

Analyses of randomised trials are often based on regression models which adjust for baseline covariates, in addition to randomised group. Based on such models, one can obtain estimates of the marginal mean outcome for the population under assignment to each treatment, by averaging the model-based predictions across the empirical distribution of the baseline covariates in the trial. We identify under what conditions such estimates are consistent, and in particular show that for canonical generalised linear models, the resulting estimates are always consistent. We show that a recently proposed variance estimator underestimates the variance of the estimator around the true marginal population mean when the baseline covariates are not fixed in repeated sampling and provide a simple adjustment to remedy this. We also describe an alternative semiparametric estimator, which is consistent even when the outcome regression model used is misspecified. The different estimators are compared through simulations and application to a recently conducted trial in asthma.