Outcome measures, placebo/nocebo magnitude, and objective sleep measurement in RLS

Synthesis of how RLS severity, augmentation, and quality of life are actually measured, how large the placebo and nocebo response is in RLS trials, and what objective (polysomnography/actigraphy) signals exist to corroborate or challenge self-reported improvement. This file exists to make every IRLS-point effect size quoted elsewhere in this knowledge base interpretable and falsifiable. All claims cite a study-note filename (in knowledge-base/studies/) or a PMID. Medical disclaimer: nothing here is individual medical advice; any change to RLS medication must be made only under physician supervision.

What is established (grade A/B)

• The International Restless Legs Scale (IRLS/IRLSSG rating scale) is the field's primary, validated 0-40 self-report severity instrument: it was validated in 196 patients and 209 controls across 20 centers in 6 countries with high internal consistency, inter-examiner reliability, test-retest reliability, and criterion validity against clinical global impression (2003-walters-irls-rating-scale-validation.md), then confirmed reliable, valid, and responsive to treatment in pooled ropinirole-trial data (2006-abetz-irls-reliability-responsiveness-rct.md) and at the item level (2005-wunderlich-irls-item-response-analysis.md). A 2014 Movement Disorder Society task force independently rated only the IRLS and the Augmentation Severity Rating Scale (ASRS) as fully "recommended" instruments among all RLS severity scales reviewed (2014-walters-severity-scales-critique-recommendations.md).
• A group-level (trial-arm) minimal clinically significant change (MCSC) for the IRLS has been estimated at approximately 3 points, derived from the IRLS-CGI-S relationship in the original validation cohort and corroborated by a meta-analytic minimal dopamine-agonist-vs-placebo difference of 3.16 points; a more conservative individual-patient threshold is approximately 6 points (2013-allen-irls-minimal-clinically-significant-change.md, PMID 24051118). An independent, larger anchor-based analysis of two 6-month rotigotine RCTs (pooled n=1,051) found a change of about -9 IRLS points corresponds to being rated "much improved" by Clinical Global Impression of Change, and about -11 points to "much/very much improved" (2016-ondo-irls-rls6-mcic-rotigotine.md, PMID 26446245).
• The placebo response in RLS drug trials is large and well quantified. The largest and most precise estimate, from 85 RCTs (5,046 participants), found a pooled placebo IRLS improvement of -6.58 points (95% CI corresponding to a standardized effect of -1.41, -1.56 to -1.25) - explicitly above the minimal clinically important difference threshold - alongside a 45.4% nocebo (adverse-event) response rate (2017-silva-placebo-nocebo-meta-analysis.md, PMID 28490647). The earlier foundational meta-analysis (36 trials) found a 40.1% placebo responder rate and confirmed the placebo effect is large for self-reported IRLS but very small for objective PLMS and essentially absent for sleep efficiency (2008-fulda-wetter-placebo-response-meta-analysis.md, PMID 17932100) - the specific Fulda & Wetter paper this KB gap-analysis was asked to check for, now confirmed present. A newer network meta-analysis restricted to alpha2delta-ligand trials corroborates a ~40% placebo CGI-I responder rate (2026-iftikhar-placebo-nocebo-network-metaanalysis.md).
• The Augmentation Severity Rating Scale (ASRS) is the validated instrument for quantifying augmentation severity (as opposed to the categorical Max Planck Institute/MPI diagnostic criteria in 2007-garcia-borreguero-max-planck-augmentation-criteria.md): in its validation study, 60% of previously untreated patients developed augmentation on levodopa, and ASRS total score clearly separated augmented from non-augmented patients (7.4 vs 2.0, p<0.0001) with good reliability (2007-garcia-borreguero-augmentation-severity-rating-scale.md, PMID 17543579).
• The Suggested Immobilization Test (SIT) is the validated objective, EMG-based provocation test for both cardinal RLS features (rest-worsening, motor restlessness): RLS patients show far higher leg discomfort and periodic-limb-movement (PLM) index than matched controls during one hour of enforced immobility (2002-michaud-suggested-immobilization-test.md, PMID 11835447).
• PSG-derived PLMI is substantially confounded by comorbid obstructive sleep apnea (OSA) under older scoring rules. The evidence-based reanalysis behind the 2016 WASM/IRLSSG scoring revision showed both AASM and WASM rules underestimate respiratory-related leg movements (thus overestimate "true" PLMS) in OSA patients (2015-manconi-respiratory-leg-movements-definition.md, already in this KB). A pediatric replication found a real OSA-PLM statistical association under 2006 scoring rules (p=0.002) that disappeared entirely under 2016 rules (p=0.677), with the proportion of children flagged as having elevated PLM roughly halving (38.9% to 19.1%) (2023-gonzalez-merino-osa-plm-children-scoring.md, PMID 38007923); a larger adult OSA cohort (n=336) confirmed WASM-2016 scoring reduces measured PLMS frequency versus the older 2006 rules (2020-schipper-plms-osa-scoring-rules.md, PMID 32858351).

Emerging C

• Night-to-night variability of PLMI is large: in RLS and PLMD patients across two consecutive PSG nights, a "Periodicity Index" (movement rhythmicity) was 2- to 6.5-fold more stable than the conventional PLMS index, meaning a single night's PLMI is a noisy outcome (2012-ferri-plmi-night-to-night-variability.md, PMID 23068780). A separate pediatric actigraphy study (non-RLS population) found PLMS varied by a mean of 16.1 events/hour across four home-monitored nights (2012-rogers-ankle-actigraphy-plms-validation.md, PMID 22841030).
• A specific validated ankle actigraph (SOMNOwatch) showed very high correlation with PSG-derived PLMI (r=0.98) and 85-93% sensitivity/specificity for detecting elevated PLMI, in a small (n=28) sleep-clinic sample (2024-spektor-somnowatch-actigraphy-plmi-validation.md, PMID 39449312) - the strongest evidence so far that a wearable could substitute for PSG in tracking objective motor activity.
• Dopaminergic pre-treatment blunts the placebo response in later trials: patients previously exposed to dopamine agonists showed smaller or even negative-slope placebo responses than drug-naive patients in two large rotigotine RCTs (2012-delafuente-fernandez-pretreatment-placebo-effect.md, PMID 22537387) - relevant because this KB's target population (long-term dopamine-agonist users) is exactly this pre-treated group.
• The IRLS correlates substantially with CGI, RLS-6, and RLSQoL, supporting IRLS plus CGI as often sufficient for trial evaluation without the full instrument battery (2013-allen-irls-cgi-rls6-qol-correlations.md, PMID 24246378); the RLSQoL/KRLS-QoL instrument itself is a separately validated, reliable quality-of-life measure (891 patients, Cronbach's alpha 0.88) already in this KB (2016-kohnen-krls-qol-validation.md).

Contested or negative findings

• Consumer wrist wearables do not measure periodic limb movements at all and have real error margins even for basic sleep staging: Fitbit Charge 2 overestimated total sleep time and light sleep and achieved only 49% accuracy for deep (N3) sleep staging, with similar limitations in a PLMS subgroup (2017-de-zambotti-fitbit-charge2-validation.md, PMID 29235907). An unrelated RCT in RLS patients found suvorexant improved self-reported sleep but produced no significant change on actigraphy-derived total sleep time or wake-after-sleep-onset - self-report and actigraphy diverged (2024-winkelman-suvorexant-actigraphy-null.md, already in this KB), underscoring that subjective improvement and objective sleep measures are not interchangeable.
• Whether OSA truly causes elevated PLM at all is now contested: manual re-scoring under 2016 rules erased a previously "significant" OSA-PLM association in children entirely (2023-gonzalez-merino-osa-plm-children-scoring.md), and a small adult re-scoring study found automated AASM software underestimates leg movements relative to manual AASM scoring, while manual AASM and manual WASM scoring were statistically similar to each other (2023-kang-plm-scoring-rules-osa.md, already in this KB) - i.e., the "right" PLMI for a given patient depends heavily on which scoring method and software were used.
• A well-designed, industry-independent pregabalin RCT (n=78) failed to beat placebo at all (adjusted 12-week IRLS change -6.8 vs -5.4, p=0.420), and the authors explicitly attributed this to a substantial placebo response comparable in magnitude to the active-drug effect (2025-sunwoo-pregabalin-korean-negative-rct.md, already in this KB) - a concrete illustration that a -6.8-point IRLS improvement, quoted elsewhere in this KB as a treatment effect, is statistically indistinguishable from the pooled placebo estimate of -6.58 points (2017-silva-placebo-nocebo-meta- analysis.md).

Mechanistic picture in 5 bullet points

• IRLS is a single-severity-factor, patient-reported 0-40 scale explaining the bulk of item variance, making it sensitive to subjective severity but, by design, unable to distinguish real symptom reduction from expectation-driven improvement (2003-walters-irls-rating-scale-validation.md).
• The RLS placebo response is domain-dependent: large for self-reported IRLS, moderate for daytime functioning, small for subjective/objective sleep parameters, and very small to absent for PLMS and sleep efficiency (2008-fulda-wetter-placebo-response-meta-analysis.md) - meaning objective PSG/PLMI outcomes are the most placebo-resistant, and self-report is the least.
• PLMI is not a fixed physiological trait measurement; it is scoring-rule-dependent and confounded by comorbid OSA, because leg movements clustering near respiratory-event termination get variably included or excluded depending on AASM vs WASM/IRLSSG rule version (2015-manconi-respiratory-leg-movements-definition.md, 2020-schipper-plms-osa-scoring-rules.md).
• PLMI (unlike the less-used Periodicity Index) has high night-to-night variability, so single-night PLMI comparisons before/after an intervention are vulnerable to regression-to-the-mean artifacts (2012-ferri-plmi-night-to-night-variability.md).
• Prior dopaminergic exposure appears to blunt subsequent placebo responsiveness (2012-delafuente-fernandez-pretreatment-placebo-effect.md), suggesting the placebo mechanism in RLS may itself route partly through the dopaminergic system this KB's target population has already had chronically modulated by years of agonist therapy.

Implications for a non-drug long-term plan for people currently on dopamine agonists or alpha-2-delta ligands

For this KB's target population - people already on a dopamine agonist or alpha-2-delta ligand - success criteria for any non-drug intervention (iron repletion, exercise, device therapy, tapering support) should be set explicitly against the placebo benchmark, not against zero: an IRLS improvement smaller than roughly 6-9 points (the anchor-based "meaningfully improved" range from 2016-ondo-irls-rls6-mcic-rotigotine.md and 2013-allen-irls-minimal-clinically-significant-change.md) should not be claimed as a real effect, and improvements in the 3-7 point range are within or near the pooled placebo response itself (2017-silva-placebo-nocebo-meta-analysis.md). Because prior dopaminergic treatment blunts placebo responsiveness (2012-delafuente-fernandez-pretreatment-placebo-effect.md), a smaller "generic" placebo bump may actually be expected in this specific population - making a genuine signal somewhat easier, not harder, to detect if a proper comparison is available. Wherever feasible, a protocol should pair self-reported IRLS with at least one objective or semi-objective corroborating signal (PSG PLMI over multiple nights, a validated ankle actigraph, or the SIT), since the placebo effect is smallest for objective PLMS and largest for subjective severity (2008-fulda-wetter-placebo-response-meta-analysis.md). Any home-tracked PLMI must be interpreted alongside comorbid-OSA status and scoring-rule version, and averaged across multiple nights rather than judged from a single reading (2012-ferri-plmi-night-to-night-variability.md, 2020-schipper-plms-osa-scoring-rules.md). Consumer wrist wearables should not be treated as RLS or PLMI monitors; at best they estimate gross sleep duration/staging with real error margins, and self-reported sleep improvement on such devices can diverge sharply from objective actigraphy (2017-de-zambotti-fitbit-charge2-validation.md, 2024-winkelman-suvorexant-actigraphy-null.md).

Open questions

• No paper identified in this search establishes a formal, peer-reviewed MCID for the RLSQoL/KRLS-QoL, the RLS-6 total score, or the ASRS - responsiveness has been shown (2016-kohnen-krls-qol-validation.md, 2016-ondo-irls-rls6-mcic-rotigotine.md item-level only) but a single agreed threshold is "not reported" for these instruments; do not invent one.
• Whether any consumer-grade wearable (smartwatch/ring) has been validated specifically for PLMI or RLS severity tracking (as opposed to general sleep staging) could not be confirmed; the closest available evidence (SOMNOwatch, a purpose-built medical actigraph, not a consumer device) suggests feasibility but does not answer the consumer-device question directly.
• Whether sham/placebo devices (e.g., inactive TENS, sham vibration, sham neuromodulation) produce a larger, smaller, or equivalent placebo response compared to placebo pills in RLS specifically was not confirmed by a dedicated head-to-head study in this search; the existing device-focused meta-analysis in this KB is restricted to alpha2delta-ligand pharmacological trials (2026-iftikhar-placebo-nocebo-network-metaanalysis.md), not device-vs-pill sham comparisons, so this question remains open.
• The true minimal clinically important difference for PSG-derived PLMI (as opposed to its reliability and OSA-related bias) does not appear to have been formally established in the literature reviewed; "not reported."

Practical instrument table

Notes in this topic

New notes added by this synthesis:

• 2003-walters-irls-rating-scale-validation.md - the landmark IRLS validation study (n=405) establishing the field's primary 0-40 severity scale.
• 2006-abetz-irls-reliability-responsiveness-rct.md - confirms IRLS reliability, validity, and responsiveness using pooled ropinirole-trial data.
• 2005-wunderlich-irls-item-response-analysis.md - item-response-theory analysis showing each IRLS item behaves near-ideally, with a minor floor effect on one item.
• 2007-garcia-borreguero-augmentation-severity-rating-scale.md - validates the ASRS, the instrument for quantifying augmentation severity.
• 2007-kohnen-rls-assessment-methodological-review.md - the single best map of the 2007-era RLS outcome-measurement landscape across severity, QoL, sleep-lab, and augmentation instruments.
• 2012-ferri-plmi-night-to-night-variability.md - shows PLMS index is far less night-to-night stable than the less-used Periodicity Index.
• 2014-walters-severity-scales-critique-recommendations.md - Movement Disorder Society task force rating only IRLS and ASRS as fully "recommended" instruments.
• 2013-allen-irls-cgi-rls6-qol-correlations.md - shows IRLS correlates substantially with CGI, RLS-6, and RLSQoL.
• 2013-allen-irls-minimal-clinically-significant-change.md - derives the ~3-point (group) / ~6-point (individual) MCID for IRLS.
• 2016-ondo-irls-rls6-mcic-rotigotine.md - anchor-based ROC analysis deriving ~9-11 point IRLS MCID thresholds against CGI-improvement categories.
• 2008-fulda-wetter-placebo-response-meta-analysis.md - foundational meta-analysis: 40% placebo responder rate, domain-dependent placebo effect size.
• 2017-silva-placebo-nocebo-meta-analysis.md - largest placebo/nocebo meta-analysis (85 RCTs): -6.58 IRLS-point pooled placebo response, 45% nocebo rate.
• 2012-delafuente-fernandez-pretreatment-placebo-effect.md - shows prior dopaminergic treatment blunts subsequent placebo response.
• 2002-michaud-suggested-immobilization-test.md - validates the SIT as an objective provocation test for RLS's cardinal features.
• 2020-schipper-plms-osa-scoring-rules.md - WASM-2016 scoring reduces measured PLMS frequency in OSA versus older 2006 rules.
• 2023-gonzalez-merino-osa-plm-children-scoring.md - a significant pediatric OSA-PLM association under 2006 rules disappears entirely under 2016 rules.
• 2024-spektor-somnowatch-actigraphy-plmi-validation.md - a medical-grade ankle actigraph shows r=0.98 correlation with PSG-derived PLMI.
• 2017-de-zambotti-fitbit-charge2-validation.md - consumer wearable (Fitbit Charge 2) does not measure PLMS and has real sleep-staging error margins.
• 2012-rogers-ankle-actigraphy-plms-validation.md - home ankle actigraphy screening is feasible but PLMS varies substantially night-to-night.

Pre-existing notes found and referenced (not rewritten):

• 2016-kohnen-krls-qol-validation.md - validates the KRLS-QoL quality-of-life instrument (891 patients).
• 2025-siriraj-immobilization-test-hrv-trial.md - recruiting trial adding heart-rate-variability analysis to the SIT in a Thai population.
• 2015-manconi-respiratory-leg-movements-definition.md - foundational evidence behind the 2016 WASM scoring-rule revision for respiratory-related leg movements.
• 2023-kang-plm-scoring-rules-osa.md - manual vs automated AASM/WASM scoring comparison in OSA patients.
• 2026-iftikhar-placebo-nocebo-network-metaanalysis.md - ~40% placebo CGI-I responder rate in alpha2delta-ligand trials specifically.
• 2007-garcia-borreguero-max-planck-augmentation-criteria.md - the categorical MPI augmentation diagnostic criteria, complementary to the ASRS's continuous severity score.
• 2025-sunwoo-pregabalin-korean-negative-rct.md - a negative RCT (-6.8 vs -5.4 IRLS points, p=0.420) illustrating that an apparently large IRLS change can be statistically indistinguishable from placebo.
• 2024-winkelman-suvorexant-actigraphy-null.md - subjective sleep improvement with no corroborating actigraphic change, illustrating self-report/objective-measure divergence.
• 2014-walters-mds-qol-instruments-review.md - MDS Task Force critique of 5 RLS quality-of-life instruments; only RLS-QOL-Abetz meets all 3 criteria for a "Recommended" rating - the interpretive anchor for translating an IRLS point-change into a quality-of-life claim (added 2026-09 quality pass, alongside 2013-allen-irls-minimal-clinically-significant-change.md and 2016-ondo-irls-rls6-mcic-rotigotine.md, which give the IRLS minimal-clinically-important-change values: a group-level MCSC of 3 points, and anchor-based MCIC cutoffs of -9 to -11 points for rotigotine trials).