Indirect Treatment Effects Research Articles

Complex Cisplatin-Double Strand Break (DSB) Lesions Directly Impair Cellular Non-Homologous End-Joining (NHEJ) Independent of Downstream Damage Response (DDR) Pathways

The treatment for advanced stage non-small cell lung cancer (NSCLC) often includes platinum-based chemotherapy and IR. Cisplatin and IR combination therapy display schedule and dose-dependent synergy, the mechanism of which is not completely understood. In a series of in vitro and cell culture assays in a NSCLC model, we investigated both the downstream and direct treatment and damage effects of cisplatin on NHEJ catalyzed repair of a DNA DSB. The results demonstrate that extracts prepared from cisplatin-treated cells are fully capable of NHEJ catalyzed repair of a DSB using a non-cisplatin-damaged DNA substrate in vitro. Similarly, using two different host cell reactivation assays, treatment of cells prior to transfection of a linear, undamaged reporter plasmid revealed no reduction in NHEJ compared with untreated cells. In contrast, transfection of a linear GFP-reporter plasmid containing site-specific, cisplatin lesions 6-bp from the termini revealed a significant impairment in DSB repair of the cisplatin-damaged DNA substrates in the absence of cellular treatment with cisplatin. Together, these data demonstrate that impaired NHEJ in combined cisplatin-IR treated cells is likely the result of a direct effect of cisplatin-DNA lesions near a DSB and that the indirect cellular effects of cisplatin treatment are not significant contributors to the synergistic cytotoxicity observed with combination cisplatin-IR treatment.

Network meta-analysis of individual and aggregate level data.

Network meta-analysis is often performed with aggregate-level data (AgD). A challenge in using AgD is that the association between a patient-level covariate and treatment effects at the study level may not reflect the individual-level effect modification. In this paper, non-linear network meta-analysis models for combining individual patient data (IPD) and AgD are presented to reduce bias and uncertainty of direct and indirect treatment effects in the presence of heterogeneity. The first method uses the same model form for IPD and AgD. With the second method, the model for AgD is obtained by integrating an underlying IPD model over the joint within-study distribution of covariates, in line with the method by Jackson et al. for ecological inferences. With simulated examples, the models are illustrated. Having IPD for a subset of studies improves estimation of treatment effects in the presence of patient-level heterogeneity. Of the two proposed non-linear models for combining IPD and AgD, the second seems less affected by bias in situations with large treatment-by-patient-level-covariate interactions, probably at the cost of greater uncertainty. Additional studies are needed to better understand when one model is favorable over the other. For network meta-analysis, it is recommended to use IPD when available. Copyright © 2012 John Wiley & Sons, Ltd.

Labeling and Cumulative Disadvantage

Research in labeling theory has been revived recently, particularly in relation to the effect of labeling on critical noncriminal outcomes that potentially exacerbate involvement in crime. This study partakes in that revitalization by examining direct and indirect effects of police intervention in the lives of adolescents who were followed into their 30s. The authors find that early police intervention is indirectly related to drug use at the ages of 29 to 31, as well as unemployment and welfare receipt. Given that such effects were found some 15 years after the labeling event, on criminal and noncriminal outcomes, and after controlling for intraindividual factors, the authors conclude that the labeling perspective is still relevant within a developmental framework.

Intact interferon signaling in peripheral blood leukocytes of high-grade osteosarcoma patients

High-grade osteosarcoma has a poor prognosis with an overall survival rate of about 60 percent. The recently closed European and American Osteosarcoma Study Group (EURAMOS)-1 trial investigates the efficacy of adjuvant chemotherapy with or without interferon-α. It is however unknown whether the interferon-signaling pathways in immune cells of osteosarcoma patients are functional. We studied the molecular and functional effects of interferon treatment on peripheral blood lymphocytes and monocytes of osteosarcoma patients, both in vivo and ex vivo. In contrast to other tumor types, in osteosarcoma, interferon signaling as determined by the phosphorylation of signal transducer and activator of transcription (STAT)1 at residue 701 was intact in immune cell subsets of 33 osteosarcoma patients as compared to 19 healthy controls. Also, cytolytic activity of interferon-α stimulated natural killer cells against allogeneic (n = 7 patients) and autologous target cells (n = 3 patients) was not impaired. Longitudinal monitoring of three osteosarcoma patients on interferon-α monotherapy revealed a relative increase in the CD16-positive subpopulation of monocytes during treatment. Since interferon signaling is intact in immune cells of osteosarcoma patients, there is a potential for indirect immunological effects of interferon-α treatment in osteosarcoma.

Partially Hidden Markov Model for Time-Varying Principal Stratification in HIV Prevention Trials

It is frequently of interest to estimate the intervention effect that adjusts for post-randomization variables in clinical trials. In the recently completed HPTN 035 trial, there is differential condom use between the three microbicide gel arms and the no-gel control arm, so intention-to-treat (ITT) analyses only assess the net treatment effect that includes the indirect treatment effect mediated through differential condom use. Various statistical methods in causal inference have been developed to adjust for post-randomization variables. We extend the principal stratification framework to time-varying behavioral variables in HIV prevention trials with a time-to-event endpoint, using a partially hidden Markov model (pHMM). We formulate the causal estimand of interest, establish assumptions that enable identifiability of the causal parameters, and develop maximum likelihood methods for estimation. Application of our model on the HPTN 035 trial reveals an interesting pattern of prevention effectiveness among different condom-use principal strata.

Interactions between human behaviour and ecological systems

Research on the interactions between human behaviour and ecological systems tends to focus on the direct effects of human activities on ecosystems, such as biodiversity loss. There is also increasing research effort directed towards ecosystem services. However, interventions to control people's use of the environment alter the incentives that natural resource users face, and therefore their decisions about resource use. The indirect effects of conservation interventions on biodiversity, modulated through human decision-making, are poorly studied but are likely to be significant and potentially counterintuitive. This is particularly so where people are dependent on multiple natural resources for their livelihoods, when both poverty and biodiversity loss are acute. An inter-disciplinary approach is required to quantify these interactions, with an understanding of human decision-making at its core; otherwise, predictions about the impacts of conservation policies may be highly misleading.

Measuring safety treatment effects using full Bayes non-linear safety performance intervention functions

Full Bayes linear intervention models have been recently proposed to conduct before–after safety studies. These models assume linear slopes to represent the time and treatment effects across the treated and comparison sites. However, the linear slope assumption can only furnish some restricted treatment profiles. To overcome this problem, a first-order autoregressive (AR1) safety performance function (SPF) that has a dynamic regression equation (known as the Koyck model) is proposed. The non-linear ‘Koyck’ model is compared to the linear intervention model in terms of inference, goodness-of-fit, and application. Both models were used in association with the Poisson-lognormal (PLN) hierarchy to evaluate the safety performance of a sample of intersections that have been improved in the Greater Vancouver area. The two models were extended by incorporating random parameters to account for the correlation between sites within comparison–treatment pairs. Another objective of the paper is to compute basic components related to the novelty effects, direct treatment effects, and indirect treatment effects and to provide simple expressions for the computation of these components in terms of the model parameters. The Koyck model is shown to furnish a wider variety of treatment profiles than those of the linear intervention model. The analysis revealed that incorporating random parameters among matched comparison–treatment pairs in the specification of SPFs can significantly improve the fit, while reducing the estimates of the extra-Poisson variation. Also, the proposed PLN Koyck model fitted the data much better than the Poisson-lognormal linear intervention (PLNI) model. The novelty effects were short lived, the indirect (through traffic volumes) treatment effects were approximately within ±10%, whereas the direct treatment effects indicated a non-significant 6.5% reduction during the after period under PLNI compared to a significant 12.3% reduction in predicted collision counts under the PLN Koyck model.

Measuring direct and indirect treatment effects using safety performance intervention functions

Recent traffic safety research has advocated the use of the Full Bayes (FB) approach to conduct controlled before–after safety evaluations. The FB approach was shown to offer both methodological and data advantages. However, there is still a lack of complete understanding of the role that the model parameters play in the formulation of treatment effectiveness. This paper offers a novel approach to compute components related to direct and indirect treatment effects under a linear intervention model for the Safety Performance Function (SPF). The isolation of a component corresponding to the direct treatment effects enables the analyst to assess the effectiveness of the countermeasures apart from local (site-related) environmental factors. Two case studies are used to demonstrate the approach. The first case study is based on the results published from a previous safety evaluation in Iowa and the second uses a new data set from British Columbia based on the Insurance Corporation of British Columbia Road Improvement Program. The various direct and indirect components have yielded valuable insight into the effects of the model parameters on treatment impacts. Moreover, the calculation of the treatment effectiveness indices have been simplified by providing straightforward equations in terms of model parameters for the computation of their components without resorting to additional algorithms. More importantly, the results can have a significant impact on the economic evaluation of safety programs and countermeasures by allowing the calculation of collision modification factors (CMFs) that vary with time.

Influence of contact pressure and boundary films on the frictional behaviour and on the roughness changes of galvanized steel sheets

This paper tackles the effects of contact pressure and of an industrial surface chemical treatment, which is called New Inorganic Treatment (NIT), on the frictional behaviour and the roughness changes of hot-dip galvanized mild steel sheets. Samples of different topographies are investigated by plane friction test with increasing contact pressure and 3D profilometry.For non-chemically treated (non-NIT) surfaces, friction coefficient is an increasing function of pressure. Moreover, above a certain pressure, which seems to depend on topography, a stick-slip phenomenon occurs. For chemically treated (NIT) surfaces, no stick-slip phenomenon is observed. Friction coefficient is lower than for non-NIT surfaces and, above a certain pressure, decreases as pressure increases. The behaviours of the 4 morphological textures investigated, either electron beam textured (EBT) or electron discharge textured (EDT), are qualitatively identical.The roughness is a decreasing function of pressure except during stick-slip phase, during which roughness evolves erratically at values higher than what the imposed pressure should lead to. The observation of height distribution curves reveals the very different surface deformations between non-NIT and NIT surfaces, whatever the topography. We conclude that, by reducing local friction shear stress, an indirect effect of the chemical treatment is the modification of the deformation mode of zinc asperities, which in fine would reduce the amount of wear debris.

Prediction of postoperative pain using path analysis in older patients

Effective postoperative pain management is important for older surgical patients because pain affects perioperative outcomes. A prospective cohort study was conducted to describe the direct and indirect effects of patient risk factors and pain treatment in explaining levels of postoperative pain in older surgical patients. We studied patients who were 65 years of age or older and were scheduled for major noncardiac surgery with a postoperative hospital stay of at least 2 days. The numeric rating scale (0 = no pain, 10 = worst possible pain) was used to measure pain levels before surgery and once daily for 2 days after surgery. Path analysis was performed to examine the association between predictive variables and postoperative pain levels. Three hundred fifty patients were studied. We found that preoperative pain level, use of preoperative opioids, female gender, higher ASA physical status, and postoperative pain control methods were the strongest predictors of postoperative pain as measured on the first day after surgery. Younger age, greater preoperative symptoms of depression, and lower cognitive function also contributed to higher postoperative pain levels. Pain levels on the second day after surgery were strongly predicted by preoperative pain level, use of preoperative opioids, surgical risk, and pain and opioid dose on postoperative day 1. However, younger age, female gender, higher ASA physical status, greater preoperative symptoms of depression, lower cognitive function, and postoperative pain control methods indirectly contributed to pain levels on the second day after surgery. Although preoperative pain and use of preoperative opioids have the strongest effects on postoperative pain, clinicians should be aware that other factors such as age, gender, surgical risk, preoperative cognitive impairment, and depression also contribute to reported postoperative pain. Based on significant statistical correlations, these study results can contribute to more effective postoperative care for those patients having the risk factors studied here. Preoperative treatment/intervention based in part on factors such as preoperative pain, use of preoperative opioids, and depression may improve postoperative pain management.

Neurotrophic Effects of Growth/Differentiation Factor 5 in a Neuronal Cell Line

The neurotrophin growth/differentiation factor 5 (GDF5) is studied as a potential therapeutic agent for Parkinson's disease as it is believed to play a role in the development and maintenance of the nigrostriatal system. Progress in understanding the effects of GDF5 on dopaminergic neurones has been hindered by the use of mixed cell populations derived from primary cultures or in vivo experiments, making it difficult to differentiate between direct and indirect effects of GDF5 treatment on neurones. In an attempt to establish an useful model to study the direct neuronal influence of GDF5, we have characterised the effects of GDF5 on a human neuronal cell line, SH-SY5Y. Our results show that GDF5 has the capability to promote neuronal but not dopaminergic differentiation. We also show that it promotes neuronal survival in vitro following a 6-hydroxydopamine insult. Our results show that application of GDF5 to SH-SY5Y cultures induces the SMAD pathway which could potentially be implicated in the intracellular transmission of GDF5's neurotrophic effects. Overall, our study shows that the SH-SY5Y neuroblastoma cell line provides an excellent neuronal model to study the neurotrophic effects of GDF5.

Benefits of Universal Intervention Effects on a Youth Protective Shield 10 Years After Baseline

An earlier randomized controlled study found that a universal, family-focused preventive intervention produced protective shield effects-reduced adolescent exposures to illicit substance opportunities-among adolescents in grade 12. This study examined a follow-up assessment of the sample during young adulthood. A randomized controlled trial evaluated the Iowa Strengthening Families Program that was implemented in 22 rural schools (N = 446 families) when the participants were in grade six. Measures included adolescent exposure to illicit substance use and young adult lifetime substance use (age 21; N = 331). Growth curve modeling examined indirect intervention effects through growth factors of adolescent exposure. Findings from this study confirm protective shield effects that mediate long-term reduction of illicit substance use (β = -.14, p = .02, Relative Reduction Rate = 28.2%). The benefits of decreasing exposure to substance use during adolescence through universal interventions were supported, with positive effects extending into young adulthood.

Analyzing direct and indirect effects of treatment using dynamic path analysis applied to data from the Swiss HIV Cohort Study

When applying survival analysis, such as Cox regression, to data from major clinical trials or other studies, often only baseline covariates are used. This is typically the case even if updated covariates are available throughout the observation period, which leaves large amounts of information unused. The main reason for this is that such time-dependent covariates often are internal to the disease process, as they are influenced by treatment, and therefore lead to confounded estimates of the treatment effect. There are, however, methods to exploit such covariate information in a useful way. We study the method of dynamic path analysis applied to data from the Swiss HIV Cohort Study. To adjust for time-dependent confounding between treatment and the outcome 'AIDS or death', we carried out the analysis on a sequence of mimicked randomized trials constructed from the original cohort data. To analyze these trials together, regular dynamic path analysis is extended to a composite analysis of weighted dynamic path models. Results using a simple path model, with one indirect effect mediated through current HIV-1 RNA level, show that most or all of the total effect go through HIV-1 RNA for the first 4 years. A similar model, but with CD4 level as mediating variable, shows a weaker indirect effect, but the results are in the same direction. There are many reasons to be cautious when drawing conclusions from estimates of direct and indirect effects. Dynamic path analysis is however a useful tool to explore underlying processes, which are ignored in regular analyses.

Modeling social, environmental and biological determinants of tuberculosis

Mathematical models have facilitated our understanding of infectious diseases dynamics and proved useful tools to compare control scenarios when interventional studies are not feasible or ethical. Here, we summarize evidence linking social, economic and biologic determinants to tuberculosis (TB) and review modeling approaches that have been used to understand their contribution to the epidemic dynamics of TB. Specifically, we find evidence for associations between smoking, indoor air pollution, diabetes mellitus, alcohol, nutritional status, crowding, migration, aging and economic trends, and the occurrence of TB infection and/or disease. We outline some methodological problems inherent to the study of these associations; these include study design issues, reverse causality and misclassification of both exposure and outcomes. We then go on to review two existing approaches to modeling the impact of determinants and the effect of interventions: the population attributable fraction model, which estimates the proportion of the TB burden that would be averted if exposure to a risk factor were eliminated from the population, and deterministic epidemic models that capture transmission dynamics and the indirect effects of interventions. We conclude by defining research priorities in both the study of specific determinants and the development of appropriate models to assess the impact of addressing these determinants.

Botulinum toxin modulates basal ganglia but not deficient somatosensory activation in orofacial dystonia

The etiology of idiopathic orofacial dystonia is incompletely understood. Neurophysiological studies indicated that a sensory dysfunction could play a key role in the pathophysiology of focal dystonia. To explore if central sensory processing is abnormal in patients with blepharospasm and Meige's syndrome and to study the effects of botulinum toxin (BTX) treatment, we systematically mapped the somatotopic representations of punctate tactile stimuli in these patients before and after therapy. Standardized tactile stimuli were pseudorandomly applied to the forehead, upper lip, and hand by a MR-compatible stimulation device during event-related fMRI. Patients showed a deficient activation in primary and secondary somatosensory representations of affected and unaffected (right hand) body regions compared to healthy controls. Although clinically effective BTX treatment did not modulate this impaired cortical activation, it reduced the activation of the thalamus and contralateral putamen during forehead stimulation. This study reveals a more generalized dysfunction of the somatosensory cortex including asymptomatic body representations in orofacial dystonia. Deficient cortical sensory activation may be due to a dedifferentiation of somatosensory representations and could represent a critical functional change within the basal ganglia-thalamocortical loops facilitating dystonic movements. Modulation of basal ganglia activation might reflect an indirect remote effect of BTX treatment on these sensorimotor circuits.

Efficacy of interspinous device versus surgical decompression in the treatment of lumbar spinal stenosis: a modified network analysis

ABSTRACT Study design: Systematic review using a modified network analysis.Objectives: To compare the effectiveness and morbidity of interspinous-device placement versus surgical decompression for the treatment of lumbar spinal stenosis.Summary: Traditionally, the most effective treatment for degenerative lumbar spinal stenosis is through surgical decompression. Recently, interspinous devices have been used in lieu of standard laminectomy.Methods: A review of the English-language literature was undertaken for articles published between 1970 and March 2010. Electronic databases and reference lists of key articles were searched to identify studies comparing surgical decompression with interspinous-device placement for the treatment of lumbar spinal stenosis. First, studies making the direct comparison (cohort or randomized trials) were searched. Second, randomized controlled trials (RCTs) comparing each treatment to conservative management were searched to allow for an indirect comparison through a modified network analysis approach. Comparison studies involving simultaneous decompression with placement of an interspinous device were not included. Studies that did not have a comparison group were not included since a treatment effect could not be calculated. Two independent reviewers assessed the strength of evidence using the GRADE criteria assessing quality, quantity, and consistency of results. The strengths of evidence for indirect comparisons were downgraded. Disagreements were resolved by consensus.Results: We identified five studies meeting our inclusion criteria. No RCTs or cohort studies were identified that made the direct comparison of interspinous-device placement with surgical decompression. For the indirect comparison, three RCTs compared surgical decompression to conservative management and two RCTs compared interspinous-device placement to conservative management. There was low evidence supporting greater treatment effects for interspinous-device placement compared to decompression for disability and pain outcomes at 12 months. There was low evidence demonstrating little to no difference in treatment effects between the groups for walking distance and complication rates.Conclusion: The indirect treatment effect for disability and pain favors the interspinous device compared to decompression. The low evidence suggests that any further research is very likely to have an important impact on the confidence in the estimate of effect and is likely to change the estimate. No significant treatment effect differences were observed for postoperative walking distance improvement or complication rates; however, findings should be considered with caution because of indirect comparisons and short follow-up periods.

Changes in health-related quality of life with smoking cessation treatment

Cigarette smoking causes reduced health-related quality of life (QoL) and smoking abstinence improves health-related QoL. We assessed the effects of treatment for tobacco dependence on the health-related QoL in a 52-week randomized controlled trial of varenicline and bupropion sustained release (SR). Subjects who smoked ≥10 cigarettes per day for the past year were randomly assigned to receive varenicline 1 mg twice daily (n = 696), bupropion SR 150 mg twice daily (n = 671) or placebo (n = 685) for 12 weeks and followed post-therapy for an additional 40 weeks. Health-related QoL was assessed using the Smoking Cessation Quality of Life questionnaire at baseline and Weeks 12, 24 and 52. Health transition (perceived health compared with baseline) and self-control were both significantly improved among subjects receiving varenicline and bupropion SR compared with placebo at Weeks 12, 24 and 52. Similarly, varenicline-treated subjects had significantly improved health transition and self-control compared with subjects who received bupropion SR at Weeks 12 and 24, and at Week 52 for health transition. A significant positive association existed between length of continuous abstinence and improved health transition, vitality, self-control, anxiety and overall mental profile. In most instances both a direct and an indirect effect (through continuous smoking abstinence) of each active treatment (vs. placebo) contributed to improved self-control and health transition. Treatment with varenicline and bupropion SR for smoking cessation resulted in improved self-control and health transition that was mediated in large part by continuous smoking abstinence.

Bioeconomic modeling of lactational antimicrobial treatment of new bovine subclinical intramammary infections caused by contagious pathogens

This study determined the direct and indirect epidemiologic and economic effects of lactational treatment of new bovine subclinical intramammary infections (IMI) caused by contagious pathogens using an existing bioeconomic model. The dynamic and stochastic model simulated the dynamics of Staphylococcus aureus, Streptococcus uberis, Streptococcus dysgalactiae, and Escherichia coli during lactation and the dry period in a 100-cow dairy herd during 1 quota year. Input parameters on cure were obtained from recent Dutch field data. The costs of clinical IMI, subclinical IMI, and intervention were calculated into the combined total annual net costs of IMI per herd. The cost effectiveness of 4 scenarios with lactational intervention was determined; scenarios included no intervention, treatment after 1 mo of infection, treatment after 2 mo of infection, and treatment after 1 mo of infection and culling of uncured cows after 2 mo of infection. Model behavior was observed for variation in parameter input values. Compared with no lactational intervention, lactational intervention of new subclinical IMI resulted in fewer clinical flare ups, less transmission within the herd, and much lower combined total annual net costs of IMI in dairy herds. Antimicrobial treatment of IMI after 1 mo of infection and culling of uncured cows after 2 mo of infection resulted in the lowest costs, whereas treatment after 2 mo of infection was associated with the highest costs between the scenarios with intervention. Changing the probability of cure resulted in a nonlinear change in the cumulative incidence of IMI cases and associated costs. Lactational treatment was able to prevent IMI epidemics in dairy herds at high transmission rates of Strep. uberis, Strep. dysgalactiae, and E. coli. Lactational treatment did not limit the spread of Staph. aureus at high transmission rates, although the associated costs were lower compared with no intervention. To improve udder health in a dairy herd, lactational treatment of contagious subclinical IMI must therefore be preceded by management measures that lower the transmission rate. Lactational treatment of environmental subclinical IMI seemed less cost effective. Detection of subclinical IMI needs improvement to be able to most effectively treat subclinical IMI caused by contagious pathogens during lactation.

Mistake on quantifying the indirect treatment effect via surrogate markers by Y. Qu and M. Case Statistics in Medicine 2006; 25:223–231

Statistics in MedicineVolume 29, Issue 19 p. 2067-2067 Letter to the Editor Mistake on quantifying the indirect treatment effect via surrogate markers by Y. Qu and M. Case Statistics in Medicine 2006; 25:223–231 Shiva Dibaj, Shiva Dibaj Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this authorSoghrat Faghihzadeh, Soghrat Faghihzadeh Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this authorShohreh Jalaie, Shohreh Jalaie Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this author Shiva Dibaj, Shiva Dibaj Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this authorSoghrat Faghihzadeh, Soghrat Faghihzadeh Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this authorShohreh Jalaie, Shohreh Jalaie Department of Biostatistics, Faculty of Medical Science, Tarbiat Modares University, Tehran, IranSearch for more papers by this author First published: 29 July 2010 https://doi.org/10.1002/sim.3932Citations: 1AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL No abstract is available for this article.Citing Literature Volume29, Issue1930 August 2010Pages 2067-2067 RelatedInformation

Pregabalin reduces sleep disturbance in patients with generalized anxiety disorder via both direct and indirect mechanisms

Background and Objectives: To characterize the impact of pregabalin on sleep in patients with generalized anxiety disorder (GAD) and to determine whether the impact is a direct or an indirect effect, mediated through the reduction of anxiety symptoms. Methods: A post-hoc analysis of data from a randomized, double-blind, placebo- and active-controlled study in patients with GAD was conducted. Patients received pregabalin 300 mg/day, venlafaxine XR 75 mg/day or placebo for a week, followed by pregabalin 300―600 mg/day, venlafaxine XR 75―225 mg/day, or placebo for 7 weeks. Treatment effect on sleep was evaluated using the Medical Outcomes Study Sleep Scale. Anxiety symptoms were assessed with the Hamilton Anxiety Rating Scale. A mediation model was used to estimate separately for both treatment arms the direct and indirect treatment effects on sleep disturbance. Results: Compared with placebo (n = 128), treatment with pregabalin (n = 121) significantly reduced scores on the sleep disturbance subscale and Sleep Problems Index II at both week 4 and week 8, and the sleep adequacy subscale at week 8. Venlafaxine XR (n = 125) had no significant effect on these measures. The mediation model indicated that 53% of the total pregabalin effect on sleep disturbance was direct (p < 0.01) and 47% indirect, mediated through anxiety symptoms (p < 0.05). Conclusions: Pregabalin decreased sleep disturbance in patients with GAD both directly, and indirectly by reducing anxiety symptoms. Given the drug specificity of the results, this study provides evidence of an additional important pathway of action for pregabalin and its efficacy in GAD.