---
canonical_name: Backward Walking
alternate_names: Retro Walking, Retrowalking, Reverse Walking, Backwards Walking, Retroambulation, Backward Locomotion, Backward Gait Training
canonical_topic: Backward Walking for Health & Longevity
short_topic_lc: backward_walking
creation_date: 2026-1024-0802
creator_ai_fullname: Opus 5.5
ep_keywords: Walking, Gait Training, Low-Impact Exercise, Aerobic Exercise
---

# Backward Walking for Health & Longevity  
<section id="top" markdown="1"></section>  
Evidence Review created on 10/24/2026 using [AI4L](https://github.com/forever-healthy/AI4L) / Opus 5.5

**Also known as:** Retro Walking, Retrowalking, Reverse Walking, Backwards Walking, Retroambulation, Backward Locomotion, Backward Gait Training

  
## Motivation

<!-- This motivation section was written only after all other sections of the document were completed, so that it reflects the full scope of the topic. -->

Backward walking, also called retro walking, is the simple act of walking in reverse, either on open ground or on a treadmill. It turns the usual step around, with the toes touching down first, the thigh and calf muscles sharing the work differently, and the path ahead out of sight, which makes an everyday movement that needs no equipment more demanding, especially for balance.

The practice comes from physical therapy, where clinicians have used it for decades in knee rehabilitation and in walking recovery after stroke. More recently it has spread as a low-cost fitness habit, promoted for knee comfort and balance, while critics question how strong the supporting studies are.

This review examines what the research shows about backward walking for adults focused on long-term health and independence: which benefits are supported and how strongly, what risks it carries, who is most likely to be affected, and how it is practiced and monitored.

**[Benefits](#expected-benefits) - [Risks](#potential-risks--side-effects) - [Protocol](#therapeutic-protocol) - [Conclusion](#conclusion)**

  
## Recommended Reading

This section lists in-depth commentary on backward walking from a science communicator, an exercise scientist, and a healthspan publication.

<!-- Search performed 2026-10-24. Web searches (WebSearch tool): "Peter Attia backward walking", "Andrew Huberman backward walking knee health", "Rhonda Patrick foundmyfitness backward walking", "Chris Kresser backward walking OR retro walking", "lifeextension.com backward walking benefits", "lifespan.io backward walking", site-restricted searches for peterattiamd.com, hubermanlab.com, foundmyfitness.com and lifeextension.com, plus "Knees Over Toes Ben Patrick backward walking" and "Janet Dufek UNLV backward walking". On-site searches: peterattiamd.com/?s=backward+walking via d-browser (21 hits, all unrelated uses of "backwards"/"walking"); hubermanlab.com/search-results?search=backward+walking via d-browser ("No results found"); foundmyfitness.com/search?q=backward walking (d-browser: Cloudflare block; d-fetch: "No results found"; d-proxy-2: 15 semantic matches, none about backward walking; the Kelly Starrett episode transcript was checked and does not discuss it); chriskresser.com/?s=backward+walking via d-fetch (unrelated podcast episodes only); lifeextension.com/search (d-browser: Access Denied; d-fetch and d-proxy-2: page shell without results; site-restricted web search: no lifeextension.com page); lifespan.io/?s=backward+walking via d-fetch ("No Articles Found"). Mainstream media (National Geographic, TIME, CNN, BBC, Fox News, AP syndications) were excluded by rule; Grokipedia, Examine and ConsumerLab are covered in their own sections; systematic reviews are in their own section. -->

- [The Benefits of Backward Walking Are Not on Solid Ground](https://www.mcgill.ca/oss/article/critical-thinking-health-and-nutrition/benefits-backward-walking-are-not-solid-ground) - Jonathan Jarry

  A science communicator's critical audit of the research, highlighting tiny trials, add-on designs that confuse direction with extra exercise, and published falls during backward-walking therapy.

- [Walking backwards has a surprising number of health benefits](https://theconversation.com/walking-backwards-has-a-surprising-number-of-health-benefits-195246) - Jack McNamara

  An exercise scientist's overview of proposed benefits, including balance, joint and back loading, energy cost, and attention demands, with links to the underlying studies.

- [Walking Backwards—Is It a Leap Forward for Healthy Aging?](https://www.insidetracker.com/a/articles/walking-backwardsis-it-a-leap-forward-for-healthy-aging) - Stevie Lyn Smith

  A healthspan-oriented summary of backward walking for balance, strength, and energy use, with practical safety precautions; published by a company that sells health testing.

Only three items are listed because no further eligible in-depth source was found; most other coverage of backward walking is mainstream media, which is excluded, or mentions it only briefly within broader interviews. No relevant content on backward walking was found from Rhonda Patrick, Peter Attia, Andrew Huberman, Chris Kresser, Life Extension Magazine, or Lifespan.io; both web and on-site searches returned only unrelated pages.

  
## Grokipedia

<!-- grokipedia.com/search?q=backward+walking was loaded with d-browser (first tier) and returned 3,598 results, the first being the dedicated article "Backward walking" (grokipedia.com/page/Backward_walking), which was then loaded with d-browser and confirmed. No further tiers were needed. -->

[Backward walking](https://grokipedia.com/page/Backward_walking)

An AI-generated encyclopedia entry covering terminology, biomechanics, claimed benefits, rehabilitation uses, and safety precautions; its claims are best checked against the primary studies.

  
## Examine

<!-- examine.com/search/?q=backward%20walking: d-browser returned a "Vercel Security Checkpoint" bot wall; d-fetch returned HTTP 429; d-proxy-1 (browser) loaded the search results, listing the dedicated page "Backward Walking" (examine.com/other/backward-walking/) and two study summaries. The dedicated page was then loaded with d-proxy-1 and confirmed. -->

[Backward Walking](https://examine.com/other/backward-walking/)

Examine's dedicated topic page summarizes physical-therapy uses for strength, gait, and joint range of motion, with a feed of study summaries on stroke and knee osteoarthritis.

  
## ConsumerLab

<!-- consumerlab.com/search/?q=backward+walking and consumerlab.com/search/?q=walking+backward were loaded with d-browser (first tier); both returned only unrelated items (multivitamins for hiking, plantar fasciitis supplements, vitamin B-12 and walking difficulty, cocoa and peripheral artery disease). No article on backward walking was listed. -->

No ConsumerLab article on backward walking exists; searches for "backward walking" and "walking backward" returned only unrelated walking-related items.

  
## Systematic Reviews

The reviews below, mostly by university physical therapy groups whose profession delivers this therapy, cover knee osteoarthritis (wear-related joint degeneration), gait impairments, balance, stroke, and body composition.

<!-- PubMed search 2026-10-24 (pubmed_search_articles): ("backward walking" OR "retro walking" OR "retro-walking" OR "backward gait" OR "retrowalking") AND (systematic review OR meta-analysis): 23 results. Selection prioritized relevance to the intervention, size, recency and coverage of distinct outcomes. Not listed: a protocol (Wu 2020), a Parkinson's systematic review without meta-analysis (Krefft 2026, cited in Emerging Research), a cerebral palsy meta-analysis (Menezes 2023), two other stroke meta-analyses (Chen 2020, Wen 2022; Chen 2020 is cited in Expected Benefits), a gait-characteristics meta-analysis (Wang 2018), a measurement-properties review of the backward walk test (Hao 2024), and reviews of broader exercise categories. No systematic review of adverse events or falls during backward walking was found. -->

- [Effectiveness of backward walking exercises combined with conventional rehabilitation programs on managing pain intensity and disability in patients with knee osteoarthritis: A systematic review and meta-analysis of randomized controlled trials](https://pubmed.ncbi.nlm.nih.gov/40924430/) - Lin et al., 2026

  Pools 13 randomized trials; adding backward walking to rehabilitation reduced knee osteoarthritis pain and disability with moderate-to-large effects.

- [The effectiveness of backward walking as a treatment for people with gait impairments: a systematic review and meta-analysis](https://pubmed.ncbi.nlm.nih.gov/30229667/) - Balasukumaran et al., 2019

  Covers 21 studies across neurological and joint conditions; benefits were clear for knee osteoarthritis, with insufficient evidence elsewhere.

- [Effectiveness of backward walking training on balance performance: A systematic review and meta-analysis](https://pubmed.ncbi.nlm.nih.gov/30616175/) - Wang et al., 2019

  Eleven studies show better balance scores and longer single-leg standing, though the authors call the evidence preliminary.

- [Backward walking training is as effective as or better than forward walking training for improving walking speed after stroke: a systematic review with meta-analysis](https://pubmed.ncbi.nlm.nih.gov/39582194/) - Menezes et al., 2025

  Separates head-to-head comparisons with forward walking from add-on designs after stroke; adding backward walking gave negligible extra speed.

- [Effectiveness of backward gait training on body composition, cardiopulmonary fitness, inflammation, and metabolic marker in adults: A systematic review and meta-analysis](https://pubmed.ncbi.nlm.nih.gov/39005976/) - Aysha et al., 2024

  Five Saudi trials in young men; backward training lowered body mass index, waist-to-height ratio, and an inflammation marker, but not body fat.

No systematic review or meta-analysis addresses the principal risk of backward walking, falls during training, so that side of the trade-off is unrepresented.

  
## Mechanism of Action

Backward walking is close to a time-reversed version of forward walking. Joint angles largely mirror the forward pattern, but muscle work flips: muscles that lengthen under load (eccentric work) during forward walking instead shorten while producing force (concentric work) during backward walking, and vice versa ([Winter et al., 1989](https://pubmed.ncbi.nlm.nih.gov/15136266/)). The forefoot lands first, the quadriceps (front-thigh muscles) become the main power producers, and the plantar flexors (calf muscles) absorb the landing.

Three effects are proposed to follow:

- **Knee unloading:** compressive force between kneecap and thigh bone (the patellofemoral joint) is lower, shown for backward running ([Flynn & Soutas-Little, 1995](https://pubmed.ncbi.nlm.nih.gov/7787851/); [Roos et al., 2012](https://pubmed.ncbi.nlm.nih.gov/22503882/)), a candidate explanation for knee-pain relief.
- **Higher energy cost:** oxygen use and heart rate are higher at any given speed ([Flynn et al., 1994](https://pubmed.ncbi.nlm.nih.gov/8133744/)), so slow speeds reach moderate intensity.
- **Sensory challenge:** without the view of the path ahead, balance relies more on proprioception (the body's sense of limb position) and the vestibular system (inner-ear balance organs), and the prefrontal cortex (the brain's planning region) works harder than during forward walking ([Kupchenko et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36940180/)).

A competing explanation holds that many reported benefits reflect added exercise time rather than direction. Most trials added backward walking on top of usual care, and one meta-analysis cautioned that its pooled effects could reflect differences in training intensity rather than walking direction ([Wang et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29804040/)). Head-to-head trials at matched volume show small or absent advantages over forward walking ([Goonasegaran et al., 2022](https://pubmed.ncbi.nlm.nih.gov/33314883/)).

  
## Historical Context & Evolution

Backward walking began as a clinical and athletic tool rather than a health trend. Popular accounts trace it to traditional Chinese exercise in public parks, but documentary support for that origin is thin. In the 1990s, sports medicine and military physical therapists used backward walking and running to keep injured athletes training while sparing the knee, and measured its oxygen cost for that purpose ([Myatt et al., 1995](https://pubmed.ncbi.nlm.nih.gov/7742838/); [Chaloupka et al., 1997](https://pubmed.ncbi.nlm.nih.gov/9130146/)). Strength coaches adopted loaded backward sled dragging, which Ben Patrick later popularized for knee conditioning; his company sells backward-walking treadmills.

From the 2000s, rehabilitation researchers tested backward walking after stroke and in Parkinson's disease, multiple sclerosis, cerebral palsy, and knee osteoarthritis. The actual findings were mostly short-term gains in walking speed, balance scores, and knee pain in trials of fewer than about 100 people. In parallel, backward walking speed emerged as a fall-risk test with good reliability ([Hao et al., 2024](https://pubmed.ncbi.nlm.nih.gov/38458049/)).

Opinion has moved in two directions. A 2000 report described two serious falls during backward-walking therapy and argued the method needs a task-specific reason and adequate safeguards ([Thomas & Fast, 2000](https://pubmed.ncbi.nlm.nih.gov/10994888/)). Later meta-analyses added pooled support in knee osteoarthritis and stroke ([Lin et al., 2026](https://pubmed.ncbi.nlm.nih.gov/40924430/); [Chen et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32629648/)), while recent critiques point out that add-on designs cannot separate walking direction from extra exercise ([Jarry, 2025](https://www.mcgill.ca/oss/article/critical-thinking-health-and-nutrition/benefits-backward-walking-are-not-solid-ground)). Head-to-head trials against forward walking and a 300-person French fall-prevention trial are the new evidence most likely to shift the picture in either direction.

  
## Expected Benefits

<!-- Benefit search 2026-10-24: PubMed (pubmed_search_articles) for ("backward walking" OR "retro walking" OR "retro-walking" OR "backward treadmill") AND (randomized OR randomised) (72 results), the systematic review query above (23 results), and targeted queries for knee osteoarthritis, stroke, multiple sclerosis, Parkinson's disease, low back pain, hamstring flexibility, body composition, cardiorespiratory responses, patellofemoral load, cognition and falls; full texts of Alghadir 2019, Aysha 2024 and Chan 2025 were read. Expert and secondary sources: McGill Office for Science and Society (Jarry 2025), The Conversation (McNamara 2022), Examine and Grokipedia topic pages. Studies finding no effect or the opposite effect were sought explicitly and are cited: Goonasegaran 2022 (no difference versus forward walking in knee osteoarthritis), the Alghadir 2019 full-text results (no significant retro-versus-forward difference, despite the abstract's wording), Kachanathu 2014 (no body-composition change), Aysha 2024 pooled body fat (no change), Menezes 2025 (negligible add-on effect after stroke), Luo 2025 (network ranking below ordinary walking), Toots 2021 and NCT04247802 (no balance gain after hip or knee replacement). A hamstring flexibility claim (Dufek, reported in press coverage) could not be traced to a PubMed-indexed study and was not graded. -->

### High 🟩 🟩 🟩

#### Knee Osteoarthritis Pain and Function ⚠️ Conflicted

In knee osteoarthritis, adding backward walking to physiotherapy reduced pain and disability across 13 randomized controlled trials (RCTs) ([Lin et al., 2026](https://pubmed.ncbi.nlm.nih.gov/40924430/)), including trials from independent groups led by [Alghadir](https://pubmed.ncbi.nlm.nih.gov/30967128/), [Gondhalekar](https://pubmed.ncbi.nlm.nih.gov/23641371/) and [Chen](https://pubmed.ncbi.nlm.nih.gov/34221078/). Alghadir's abstract implies superiority over forward walking, but its full-text results show no significant backward-versus-forward difference; a 34-person RCT likewise found none ([Goonasegaran et al., 2022](https://pubmed.ncbi.nlm.nih.gov/33314883/)). Net: backward walking clearly adds to physiotherapy alone, but its advantage over equal forward walking is unproven.

**Magnitude:** Versus control programs, pain fell by a Hedges' g (standardized effect size in standard-deviation units) of −1.00 (95% confidence interval [CI], the range likely to contain the true effect, −1.37 to −0.62) and disability by −1.02 (95% CI −1.33 to −0.70) ([Lin et al., 2026](https://pubmed.ncbi.nlm.nih.gov/40924430/)).

#### Balance and Walking Speed ⚠️ Conflicted

Balance stability scores and single-leg standing improved versus control across 11 studies ([Wang et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30616175/)), and adding backward walking to usual stroke care raised the Berg Balance Scale (a 0–56 clinical balance test) across 4 trials ([Chen et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32629648/)), including trials led by [Yang](https://pubmed.ncbi.nlm.nih.gov/15859527/) and [Rose](https://pubmed.ncbi.nlm.nih.gov/29232308/). An earlier meta-analysis found no significant neurological benefit ([Balasukumaran et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30229667/)), a joint-replacement pilot found none ([NCT04247802](https://clinicaltrials.gov/study/NCT04247802)), and adding backward to forward training gave negligible extra speed ([Menezes et al., 2025](https://pubmed.ncbi.nlm.nih.gov/39582194/)). Net: balance gains are supported but small and inconsistent.

**Magnitude:** Berg Balance Scale 3.82 points higher than control (95% CI 2.12 to 5.52) ([Chen et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32629648/)); walking speed 0.16 m/s faster than forward training (95% CI 0.06 to 0.27), but only 0.03 m/s (95% CI 0.01 to 0.04) when added to forward training ([Menezes et al., 2025](https://pubmed.ncbi.nlm.nih.gov/39582194/)).

### Medium 🟩 🟩

#### Lower-Limb Strength and Power

In knee osteoarthritis, quadriceps strength rose more with six weeks of level backward walking plus physiotherapy than with physiotherapy alone ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)). In recreational athletes, six weeks of backward walking and running on inclines improved leg-press strength and jump power within the training group, against controls who kept their normal training, without a reported between-group test ([Ferley et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36455596/)); that evidence is indirect for level walking because it used inclines and running. The mechanism is greater concentric work by the knee extensors.

**Magnitude:** Quadriceps strength rose 1.0 kg more than in the control group (1.7 versus 0.7 kg; p = 0.008, where the p-value is the probability of a chance finding) ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); Ferley reported figures for the training group only (no control-group figure), so its gains are not compared between groups.

### Low 🟩

#### Cardiorespiratory Fitness at Low Speeds

Backward walking demands more oxygen and heart rate than equal-speed forward walking ([Flynn et al., 1994](https://pubmed.ncbi.nlm.nih.gov/8133744/); [Chaloupka et al., 1997](https://pubmed.ncbi.nlm.nih.gov/9130146/)). Training evidence is indirect: a non-randomized trial used walk–run training ([Terblanche et al., 2005](https://pubmed.ncbi.nlm.nih.gov/15776337/)); per a meta-analysis, the randomized trial showing aerobic gains used inclines ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)).

**Magnitude:** Predicted aerobic capacity rose 5.2% within the training group (p = 0.013); the abstract gives no control-group figure, so the result is not compared between groups ([Terblanche et al., 2005](https://pubmed.ncbi.nlm.nih.gov/15776337/)). Backward walking at 3.5 mph (5.6 km/h) costs 6.0 METs (metabolic equivalents, multiples of resting energy use) versus 4.8 METs for brisk forward walking, per the [2024 Adult Compendium of Physical Activities](https://pubmed.ncbi.nlm.nih.gov/38242596/).

#### Body Composition and Cardiometabolic Markers ⚠️ Conflicted

In subsets of five Saudi RCTs in men, 10%-incline backward training, indirect for level walking, lowered body mass index (BMI), waist-to-height ratio and C-reactive protein (CRP, inflammation marker) versus forward training ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)). Body fat did not differ ([Kachanathu et al., 2014](https://pubmed.ncbi.nlm.nih.gov/25540487/)). Net: marker gains lack body-fat confirmation.

**Magnitude:** Versus forward training, standardized mean differences (effect sizes in standard-deviation units) were −0.55 for BMI (95% CI −0.77 to −0.32), −1.18 for waist-to-height ratio (95% CI −1.89 to −0.48), −0.98 for CRP (95% CI −1.28 to −0.70), and −0.11 for body-fat percentage (95% CI −0.62 to 0.40) ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)).

#### Low Back Pain Relief

In 30 adults with chronic low back pain worsened by bending, one 15-minute backward treadmill session reduced pain and improved lower-back movement control ([Chan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40901866/)). Forward walking reduced pain similarly, with no between-group difference; no multi-week trial exists. Proposed mechanism: greater back-extensor recruitment.

**Magnitude:** Pain (0–10 numerical rating scale) fell from 3.69 to 2.75 after backward walking (p = 0.014) and from 2.86 to 1.79 after forward walking (p = 0.026, not significant after correction for multiple tests); the change did not differ between groups (time-by-group interaction p = 0.45) ([Chan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40901866/)).

### Speculative 🟨

#### Fall Prevention in Healthy Older Adults

Slower backward walking predicts later falls in nursing-home residents ([Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/)), suggesting training it could protect. No trial has yet reported falls after training; the basis is mechanistic only.

#### Patellofemoral Joint Unloading

Backward running produced lower kneecap–thigh-bone compression than forward running in two biomechanical studies ([Flynn & Soutas-Little, 1995](https://pubmed.ncbi.nlm.nih.gov/7787851/); [Roos et al., 2012](https://pubmed.ncbi.nlm.nih.gov/22503882/)). No controlled trial in kneecap pain exists; the basis is mechanistic only.

#### Cognitive Engagement

Backward walking raised prefrontal brain activity above forward walking in a small observational study ([Kupchenko et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36940180/)). No trial shows lasting cognitive benefit from training; the basis is mechanistic only.

  
## Benefit-Modifying Factors

- **Genetic polymorphisms:** No gene variant has been studied as a modifier of response to backward walking; general exercise-response genetics have not been tested for this form of training.
- **Baseline walking ability:** After stroke, people with severe baseline walking impairment (0.4 m/s or slower) gained less overall from backward treadmill training than milder cases ([Awosika et al., 2022](https://pubmed.ncbi.nlm.nih.gov/35203897/)).
- **Baseline body composition:** BMI and CRP reductions were reported in overweight or obese young men; trials in normal-weight men found no body-composition change ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)).
- **Sex:** The body-composition trials enrolled only men, while the only women's training study reported fat loss ([Terblanche et al., 2005](https://pubmed.ncbi.nlm.nih.gov/15776337/)); no trial compares responses by sex.
- **Pre-existing conditions:** Knee trials enrolled Kellgren-Lawrence grades 1–3 (a 0–4 X-ray severity scale) ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); after hip or knee replacement, a 63-person pilot (42 completers) found no balance gain (Berg 53.0 versus 52.4; [NCT04247802](https://clinicaltrials.gov/study/NCT04247802)).
- **Age:** Knee trials enrolled ages 45–66 ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); community-dwelling adults averaging 76 years walked backward at about half their forward speed ([Gong et al., 2026](https://pubmed.ncbi.nlm.nih.gov/41703240/)), so starting speeds and achievable intensity differ.
- **Comparison activity:** Gains over usual care are clearer than over matched forward walking; a network meta-analysis (one comparing many treatments at once) ranked retro walking seventh of ten aerobic options for knee osteoarthritis ([Luo et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40452037/)).

  
## Potential Risks & Side Effects

<!-- Risk search 2026-10-24: no prescribing information exists for an exercise, so a clinical reference source was used instead: the Cleveland Clinic Health Essentials page on walking backwards (health.clevelandclinic.org/benefits-of-walking-backwards; safety advice to start slowly, use handrails and a clear area), together with the McGill Office for Science and Society review (which cites the Thomas & Fast 2000 case report). PubMed (pubmed_search_articles): (backward walking OR retro walking OR backward treadmill) AND (adverse events OR safety OR falls OR injury OR "muscle soreness" OR "muscle damage") (336 results screened), "backward downhill walking" (23 results), and dizziness, Parkinson's disease and falls queries. Full texts of Alghadir 2019 and Awosika 2020 were read for adverse events; ClinicalTrials.gov results for NCT04247802 were retrieved. Studies reporting no excess or lower risk were sought and are cited: Alghadir 2019 (no serious adverse events in any arm; mild pain flares only in the forward-walking arm), DelMastro 2023 (no safety concerns), Awosika 2020 (no serious adverse events), Chan 2025 (no adverse responses), Hammer 2021 (no adverse events with backward downhill running in Achilles tendinopathy). -->

### High 🟥 🟥 🟥

No risk reaches High: harm data consist of two case reports, laboratory exercise studies, and small trials that were not designed to compare adverse-event rates between groups.

### Medium 🟥 🟥

#### Delayed-Onset Muscle Soreness and Transient Strength Loss

Walking backward down a slope loads the calf and thigh muscles while they lengthen (eccentric contraction), causing soreness and temporary weakness. Independent groups led by [Nottle](https://pubmed.ncbi.nlm.nih.gov/16248467/) and [Hoffman](https://pubmed.ncbi.nlm.nih.gov/23558392/) (Australia) and [Girard](https://pubmed.ncbi.nlm.nih.gov/30016189/) (Qatar and France) documented this after 30–60 minutes at −15% to −25% grades. The gap: these protocols are much harder than level backward walking, so the risk applies mainly to steep, long, or unaccustomed sessions. Soreness peaks around 48 hours and resolves within days.

**Magnitude:** Calf (plantar-flexor) torque fell 18.1% after backward downhill walking versus 6.0% (non-significant) after distance-matched forward uphill walking, within-condition changes not compared between groups, and soreness rose only after the backward downhill bout ([Girard et al., 2018](https://pubmed.ncbi.nlm.nih.gov/30016189/)).

### Low 🟥

#### Falls and Fall-Related Injury During Training ⚠️ Conflicted

Walking without seeing the path raises tripping risk; two patients fell during backward-walking therapy with considerable injury ([Thomas & Fast, 2000](https://pubmed.ncbi.nlm.nih.gov/10994888/)). Supervised RCTs reported no serious adverse events ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/); [DelMastro et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36923577/)). Net: supervised risk appears low; unsupervised outdoor risk is unmeasured.

**Magnitude:** Serious adverse events numbered zero in the backward-walking, forward-walking, and control arms of the 68-person Alghadir trial, which did not report falls separately ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); the case reports have no comparison group.

### Speculative 🟨

#### Excess Cardiovascular Strain

Backward walking raises heart rate above forward walking at equal speed ([Flynn et al., 1994](https://pubmed.ncbi.nlm.nih.gov/8133744/)), possibly overshooting intended intensity in heart disease. No cardiac events are reported; the basis is mechanistic only.

  
## Risk-Modifying Factors

- **Genetic polymorphisms:** No gene variant has been identified that modifies fall, soreness, or other risks during backward walking.
- **Baseline balance:** A 3-meter backward walk time of 8.15 seconds or more marked 27% higher prospective fall incidence in nursing-home residents ([Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/)), flagging people who face higher fall risk during unsupported practice.
- **Sex:** No sex-based difference in adverse events has been reported; most body-composition trials enrolled only men, so harm data in women are especially sparse.
- **Pre-existing conditions:** Parkinson's disease magnifies backward-walking deficits ([Hackney & Earhart, 2009](https://pubmed.ncbi.nlm.nih.gov/18951535/)), and people with dizziness walk backward more slowly ([Davalos-Bichara et al., 2014](https://pubmed.ncbi.nlm.nih.gov/25042815/)); after hip or knee replacement, 1 of 22 backward-walking participants had a postoperative joint dislocation versus 0 of 20 controls ([NCT04247802](https://clinicaltrials.gov/study/NCT04247802)).
- **Age:** Both published therapy falls involved older patients, one with a hip fracture ([Thomas & Fast, 2000](https://pubmed.ncbi.nlm.nih.gov/10994888/), as summarized by [Jarry](https://www.mcgill.ca/oss/article/critical-thinking-health-and-nutrition/benefits-backward-walking-are-not-solid-ground)), and older adults at high fall risk walk backward about 25% slower than healthy peers ([Yada et al., 2021](https://pubmed.ncbi.nlm.nih.gov/34880541/)).
- **Training history:** Soreness and strength loss appeared after a first backward downhill bout but not after a repeat bout one week later ([Hoffman et al., 2016](https://pubmed.ncbi.nlm.nih.gov/27433958/)), so novices face the most soreness.

  
## Key Interactions & Contraindications

<!-- Interaction search 2026-10-24: PubMed (pubmed_search_articles) for backward walking with "drug interaction" OR "pharmacokinetics" returned no human interaction or pharmacokinetic study of backward walking; none is expected for an exercise. Interactions were therefore derived from human evidence on drugs that raise fall risk in general (EUGMS systematic reviews: Seppala 2018 psychotropics and others; de Vries 2018 cardiovascular drugs), the 2023 AGS Beers Criteria, and the American Diabetes Association exercise position statement (Colberg 2016). Because none of these studied backward walking itself, every interaction below is marked theoretical. -->

- **Psychotropic drugs (acting on mood, anxiety or thought) (diazepam, alprazolam; sertraline, amitriptyline; quetiapine):** Caution (theoretical). These are linked to 1.3- to 2.0-fold fall odds in older adults ([Seppala et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29402652/)), likely compounding backward-walking instability. Mitigation: handrail-supported treadmill, supervision, and sessions away from peak sedation.
- **Opioids (potent pain-relieving drugs) and antiepileptics (seizure and nerve-pain drugs) (oxycodone, tramadol; gabapentin, pregabalin):** Caution (theoretical). Their fall odds ratios (OR, relative odds versus non-users) were 1.60 and 1.55 ([Seppala et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29402646/)); dizziness may destabilize backward steps. Mitigation: rail-supported sessions.
- **Loop diuretics (drugs that increase urine output) and blood-pressure-lowering drugs (furosemide; amlodipine, lisinopril):** Caution (theoretical). Loop diuretics carry a fall OR of 1.36 ([de Vries et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29396189/)); light-headedness on stopping or turning may cause a backward fall. Mitigation: hydration and a pause before turning.
- **Beta-blockers (heart-rate-lowering drugs) (metoprolol, atenolol):** Monitor (theoretical). They blunt heart-rate rise, so heart-rate targets understate the higher effort of backward walking; their fall OR was 0.88 ([de Vries et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29396189/)). Mitigation: intensity gauged by perceived exertion.
- **Insulin and sulfonylureas (oral insulin-releasing drugs) (insulin glargine; glipizide, glimepiride):** Monitor (theoretical). Exercise raises hypoglycemia (low blood sugar) risk with these drugs, per an American Diabetes Association statement whose members gain no revenue from it ([Colberg et al., 2016](https://pubmed.ncbi.nlm.nih.gov/27926890/)). Mitigation: a pre-session glucose check.
- **Anticoagulants (drugs that slow blood clotting) (warfarin, apixaban, rivaroxaban):** Caution (theoretical). No interaction study exists; a backward fall onto the head or hip carries higher bleeding risk on these drugs, including bleeding inside the skull. Mitigation: handrail-supported practice or soft, clear surfaces.
- **Over-the-counter sedating antihistamines (drowsiness-causing allergy drugs) (diphenhydramine, doxylamine):** Caution (theoretical). The [2023 American Geriatrics Society Beers Criteria](https://pubmed.ncbi.nlm.nih.gov/37139824/) list them as potentially inappropriate for older adults because of sedation and confusion; its members earn no revenue from this advice. Mitigation: sessions spaced hours after a dose.
- **Over-the-counter anti-inflammatory pain relievers (ibuprofen, naproxen):** Monitor (theoretical). These anti-inflammatory drugs showed no significant fall association (OR 1.09, 95% CI 0.96–1.23; [Seppala et al., 2018](https://pubmed.ncbi.nlm.nih.gov/29402646/)), but masking pain may hide early overuse soreness. Mitigation: progression by pain-free sessions.
- **Sedating supplements (valerian, kava, melatonin):** Caution (theoretical). No interaction study exists; additive drowsiness may impair balance while walking backward. Mitigation: no backward walking after evening doses.
- **Supplements with additive effects (berberine, chromium; beetroot nitrate):** Monitor (theoretical). Glucose-lowering and blood-pressure-lowering supplements may add to exercise-related drops, especially alongside medication; no interaction study exists. Mitigation: glucose or blood-pressure checks when combining.
- **Other interventions:** Monitor (theoretical). Combining backward treadmill training with spinal electrical stimulation gave no extra walking gain ([Awosika et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32954299/)). Same-day heavy leg training adds eccentric load and soreness; handrails lower effort.

**Populations who should avoid Backward Walking:**

- People with recent falls or severe balance impairment (for example, a 3-meter backward walk time of 8.15 seconds or more; [Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/)), unless supervised with parallel bars, rails, or a harness ([Thomas & Fast, 2000](https://pubmed.ncbi.nlm.nih.gov/10994888/))
- People in early recovery from hip or knee replacement without surgical clearance, given one postoperative joint dislocation (1 of 22 backward-walking participants versus 0 of 20 controls) in the only joint-replacement trial ([NCT04247802](https://clinicaltrials.gov/study/NCT04247802))
- People with active vertigo (a spinning sensation) or an acute vestibular disorder
- People with unstable cardiac symptoms such as exertional chest pain or unexplained breathlessness, New York Heart Association class III–IV heart disease (symptoms with ordinary or minimal activity; excluded in [NCT04247802](https://clinicaltrials.gov/study/NCT04247802)), or a heart attack or heart surgery within 3 months (excluded in [NCT04928482](https://clinicaltrials.gov/study/NCT04928482)), until medically evaluated

  
## Risk Mitigation Strategies

Doses, speeds, and timings below follow common practice unless cited.

- **Treadmill with handrails first:** Learning typically starts on a treadmill at about 1.5–2.5 km/h with hands on the rails before overground practice; prevents learning-phase falls, as parallel bars or suspension support were recommended after therapy falls ([Thomas & Fast, 2000](https://pubmed.ncbi.nlm.nih.gov/10994888/)).
- **Clear, flat path outdoors:** Overground practice uses an empty track, field, or hallway, with glances over a shoulder or a spotter; prevents tripping and collisions with obstacles or people.
- **Short initial bouts:** Programs start with 10 minutes, three times weekly, progressing toward 30 minutes over six weeks ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); limits soreness and knee-pain flares.
- **Pain-guided progression:** Duration increases only while pain stays below 3 on a 0–10 scale ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); prevents aggravating osteoarthritic knees.
- **Stop threshold for pain:** Sessions are reduced or paused if knee pain increases for 2 days or more, the trial's adverse-effect definition ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); prevents overuse injury.
- **Gentle grades before downhill work:** Early sessions stay level, with declines added only after several level sessions, because a first steep bout causes most soreness and later bouts much less ([Hoffman et al., 2016](https://pubmed.ncbi.nlm.nih.gov/27433958/)); mitigates muscle soreness and strength loss.
- **Effort by perceived exertion:** Speed is set by breathing and effort rather than familiar forward-walking speeds; prevents cardiovascular overshoot from backward walking's higher heart-rate cost.
- **Medication timing:** Sessions are scheduled away from sedating doses, with a glucose check for insulin or sulfonylurea users; reduces dizziness- and hypoglycemia-related falls.
- **Stop for warning symptoms:** Sessions end at dizziness, chest pain, or unusual breathlessness; prevents falls and cardiac events.

  
## Therapeutic Protocol

Timing, cycling, progression steps, and session structure below reflect common practice unless cited.

- **Knee osteoarthritis regimen:** 10 minutes of supervised backward walking at comfortable speed, three times weekly for six weeks, progressing toward 30 minutes, added to physiotherapy ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)).
- **Speed and surface:** Subgroup analyses found larger knee effects at comfortable walking speed and with treadmill walking ([Lin et al., 2026](https://pubmed.ncbi.nlm.nih.gov/40924430/)).
- **Body-composition regimen:** Treadmill training four times weekly for 12 weeks in overweight and obese young adults ([Soman et al., 2023](https://pubmed.ncbi.nlm.nih.gov/37008812/)); included trials used 4 km/h or 4 mph at 10% incline ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)).
- **Neurological rehabilitation regimen:** 30-minute sessions three times weekly, for two weeks after stroke ([Awosika et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32954299/)) or eight weeks in multiple sclerosis ([DelMastro et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36923577/)).
- **Home micro-dose approach:** 5 minutes daily, progressing to at least 10 minutes daily, within a 12-week home program after joint replacement ([NCT04247802](https://clinicaltrials.gov/study/NCT04247802)).
- **Loaded approach:** Backward sled dragging, popularized by Ben Patrick, has no trial-defined dose ([Ferriss podcast](https://tim.blog/2025/11/11/ben-patrick-kneesovertoesguy/)); incline backward walking and running three times weekly is the closest tested regimen ([Ferley et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36455596/)).
- **Clinic versus fitness approaches:** Physiotherapists use backward walking as a supervised add-on for specific conditions; fitness practitioners use it as unsupervised cross-training. Neither approach has been compared with the other.
- **Time of day:** No study has compared timing; sessions fit any time, avoiding periods of fatigue, sedation, or low blood glucose.
- **Half-life:** Not applicable; backward walking is an exercise, not a compound, and its effects depend on accumulated sessions rather than blood levels.
- **Single or split sessions:** No study compares one session with several shorter bouts; trials used one session of 10–30 minutes per training day ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/); [DelMastro et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36923577/)).
- **Genetic polymorphisms:** No variant is known to influence the choice of protocol or training dose.
- **Sex:** Speed-to-oxygen-uptake curves exist separately for women ([Clarkson et al., 1997](https://pubmed.ncbi.nlm.nih.gov/9007769/)) and men ([Myatt et al., 1995](https://pubmed.ncbi.nlm.nih.gov/7742838/)); no sex-specific dosing data exist.
- **Age:** Adults averaging 76 years chose a backward speed of about 0.73 m/s ([Gong et al., 2026](https://pubmed.ncbi.nlm.nih.gov/41703240/)), roughly half forward speed; older starters begin slower and on rails.
- **Baseline balance markers:** A 3-meter backward walk time of 8.15 seconds or more ([Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/)) points to supervised, rail-supported practice only.
- **Pre-existing conditions:** Evidence supports use in knee osteoarthritis grades 1–3, stroke, and multiple sclerosis under supervision; after joint replacement, programs run through physiotherapists.

  
## Discontinuation & Cycling

- **Lifelong or short-term:** Trials lasted 3–12 weeks with little follow-up; continued use as one part of varied training is common practice, but long-term effects are unstudied.
- **Withdrawal effects:** None are known; as with other training, gains are expected to fade gradually after stopping, and maintenance beyond the intervention period remains uncertain ([Menezes et al., 2025](https://pubmed.ncbi.nlm.nih.gov/39582194/)).
- **Tapering:** No tapering is needed; sessions can stop at any time without physiological rebound.
- **Cycling:** No evidence supports cycling to maintain efficacy; after a break, restarting at shorter, level sessions avoids renewed soreness, since protection from repeat bouts fades with time off.

  
## Sourcing and Quality

- **Largely not applicable:** Backward walking needs no supplement, prescription, or product; quality considerations apply only to equipment, footwear, and setting.
- **Motorized treadmills:** Relevant features are full-length handrails, fine speed increments starting below 1 km/h, and an emergency-stop safety key; neurological trials used rails or harness support ([Awosika et al., 2020](https://pubmed.ncbi.nlm.nih.gov/32954299/)).
- **Non-motorized backward treadmills and sleds:** Products such as the ATG Backward/Forward Resisted Treadmill and push-pull sleds provide resisted backward walking; no independent testing exists, and the maker has a financial interest in their adoption.
- **Footwear:** Flat, flexible shoes with good grip suit the forefoot-first landing; thick heel cushioning offers no advantage when the heel lands last.
- **Surface:** Running tracks, sports fields, and empty corridors give firm, predictable footing; uneven trails, curbs, and traffic add trip and collision hazards.

  
## Practical Considerations

- **Time to effect:** Knee-pain and disability improvements appeared after 3 weeks ([Gondhalekar & Deo, 2013](https://pubmed.ncbi.nlm.nih.gov/23641371/)) to 6 weeks ([Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/)); body-composition marker changes were measured after 12 weeks ([Soman et al., 2023](https://pubmed.ncbi.nlm.nih.gov/37008812/)).
- **Common pitfalls:** Starting outdoors or at forward-walking speeds, twisting the neck repeatedly to look back, adding steep declines too early, and treating backward walking as a substitute for overall activity volume.
- **Regulatory status:** Backward walking is an unregulated exercise technique; it has no approved indication and is delivered within physiotherapy or fitness practice.
- **Cost and accessibility:** Backward walking is free and needs no equipment. Because no manufacturer profits from it, large trials lack industry funding; payers may favor such low-cost exercise over costlier treatments, a possible structural bias in guidance.
- **Learning curve:** Backward walking speed is roughly half of forward speed in older adults ([Gong et al., 2026](https://pubmed.ncbi.nlm.nih.gov/41703240/)); most people need several sessions before moving comfortably.
- **Activity trackers:** Wrist devices estimate calories with forward-walking assumptions, so they likely underestimate backward walking's higher energy cost (theoretical).

  
## Interaction with Foundational Habits

- **Sleep:** None known (indirect at most). No study has tested backward walking on sleep; as moderate exercise it may support sleep like other walking. Practical consideration: demanding treadmill sessions earlier in the day avoid delayed sleep onset in people sensitive to late exercise.
- **Nutrition:** Indirect. Backward walking burns more energy per minute than forward walking at the same speed (6.0 versus 4.8 METs; [Herrmann et al., 2024](https://pubmed.ncbi.nlm.nih.gov/38242596/)); trials in overweight men did not control diet ([Aysha et al., 2024](https://pubmed.ncbi.nlm.nih.gov/39005976/)). Practical consideration: adequate protein after eccentric-heavy downhill sessions may support muscle recovery.
- **Exercise:** Potentiating and complementary. It adds quadriceps and calf work with lower kneecap compression, suiting warm-ups and cross-training. Blunting risk: steep backward downhill walking cut calf strength for 72 hours ([Girard et al., 2018](https://pubmed.ncbi.nlm.nih.gov/30016189/)), so it is best kept away from heavy leg-training days.
- **Stress management:** None known (indirect). Backward walking requires focused attention on body position, which may feel absorbing; no study has measured cortisol or stress responses. Practical consideration: a quiet, safe space allows attention to stay on movement rather than hazards.

  
## Monitoring Protocol & Defining Success

Baseline testing before starting covers three areas: balance and fall risk, using the 3-meter backward walk test and a count of falls in the past year; the target condition, using a knee questionnaire for osteoarthritis or waist-to-height ratio for body composition; and cardiovascular readiness, including resting heart rate and a review of medications that sedate or lower blood pressure or glucose. People with exertional chest pain or unexplained breathlessness obtain medical clearance first.

Ongoing monitoring follows this cadence: heart rate and perceived effort during every session for the first 2 weeks; pain scores after each session and the next day; the backward walk test and knee questionnaire at 3 and 6 weeks, matching trial lengths; waist-to-height ratio at 12 weeks; then all measures every 3–6 months while practice continues.

| Biomarker | Optimal Functional Range | Why Measure It? | Context/Notes |
|-----------|--------------------------|-----------------|---------------|
| 3MBWT time | Under 8.15 seconds (prospective-fall cutoff; Kapan et al., 2025) | Safety check: slower times limit practice to supervised, rail-supported sessions | 3MBWT: 3-meter backward walk test, time to walk 3 m backward at usual pace. Cutoff from [Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/) in nursing-home residents. Conventional reference values in healthy adults (25th–75th percentile): 2.23–3.00 seconds in men and 2.60–3.50 seconds in women ([Mbada et al., 2023](https://pubmed.ncbi.nlm.nih.gov/37852188/)). Same footwear and floor each time. |
| Exercise heart rate | 64–76% of maximal heart rate for moderate intensity (ACSM; Garber et al., 2011) | Safety check: backward walking raises heart rate above forward walking at the same speed | ACSM: American College of Sports Medicine; its members include exercise professionals who may earn revenue from exercise prescription. Source: [Garber et al., 2011](https://pubmed.ncbi.nlm.nih.gov/21694556/). Conventional resting heart rate 60–100 beats per minute. Beta-blockers blunt readings. |
| WOMAC score (knee osteoarthritis only) | No established target; track change from own baseline | Expected to change: knee pain and function | WOMAC: Western Ontario and McMaster Universities Osteoarthritis Index, a validated knee questionnaire on which lower scores are better. Complete at the same time of day each time. |
| Waist-to-height ratio | Under 0.5 (Ashwell et al., 2012) | Expected to change: central body fat | Screening boundary from the meta-analysis by [Ashwell et al., 2012](https://pubmed.ncbi.nlm.nih.gov/22106927/); no conventional laboratory reference range exists. Measure the waist midway between lowest rib and hip bone after a normal exhale; fasting not required. |
| Pain score (0–10 numerical rating scale) | No established target; track change from own baseline | Safety check: pain rising for 2 days or more pauses progression | The 2-day rule follows the adverse-effect definition of [Alghadir et al., 2019](https://pubmed.ncbi.nlm.nih.gov/30967128/). Record after each session and the next morning. |

Qualitative markers:

- **Balance confidence:** Comfort stepping backward away from a counter or curb, and steadiness when turning
- **Knee comfort:** Ease on stairs and when rising from a chair
- **Soreness pattern:** Calf or thigh soreness 24–72 hours after sessions, expected to fade with repetition
- **Walking ease:** Perceived effort and stability during ordinary forward walking
- **Attention and focus:** Mental fatigue after sessions and the ability to stay oriented without frequent glances backward

  
## Emerging Research

- **Fall prevention after age 75:** A French randomized trial ([NCT07650097](https://clinicaltrials.gov/study/NCT07650097); 300 participants; not yet recruiting) swaps part of day-hospital rehabilitation for five weeks of supervised backward walking at equal total time; primary endpoint: falls over 12 months. A positive result would raise fall prevention above Speculative; a null result would weaken the balance rationale.
- **Multiple sclerosis (TRAIN-BW):** Backward versus forward walking training for 8 weeks ([NCT04091464](https://clinicaltrials.gov/study/NCT04091464); 90 participants; primary endpoints postural sway and stepping latency); completed May 2026, no results posted ([protocol, VanNostrand et al., 2024](https://pubmed.ncbi.nlm.nih.gov/38971304/)). Results will show whether balance gains exceed matched forward walking.
- **Dose after stroke:** A Veterans Affairs trial compares 18 with 27 sessions and uses brain imaging to predict response ([NCT04928482](https://clinicaltrials.gov/study/NCT04928482); 48 participants; recruiting; completion listed December 2026). A dose effect would support longer programs; a null result would favor shorter ones.
- **Older nursing-home residents:** A 52-participant randomized trial of backward walking on balance, fear of falling, and fall incidence ([NCT05409534](https://clinicaltrials.gov/study/NCT05409534)) completed in 2024; no results posted.
- **Chronic low back pain:** A 42-participant trial of added retro walking ([NCT07484542](https://clinicaltrials.gov/study/NCT07484542)) completed in June 2026; no results posted. It tests over six weeks the pain signal seen after a single session.
- **Direction versus volume:** A network meta-analysis ranked retro walking seventh of ten aerobic options for knee osteoarthritis ([Luo et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40452037/)), and a 34-person trial found no advantage over forward walking ([Goonasegaran et al., 2022](https://pubmed.ncbi.nlm.nih.gov/33314883/)); larger volume-matched trials could weaken the case.
- **Parkinson's disease:** A 2026 systematic review found only two randomized trials and called the evidence preliminary ([Krefft et al., 2026](https://pubmed.ncbi.nlm.nih.gov/42195120/)); new trials could strengthen or weaken the neurological case.
- **Backward walking as a screening test:** The 3-meter backward walk test predicted prospective falls as well as a standard physical performance battery ([Kapan et al., 2025](https://pubmed.ncbi.nlm.nih.gov/40187370/)), and a 2026 scoping review supports its validity in neurological disorders ([VanNostrand et al., 2026](https://pubmed.ncbi.nlm.nih.gov/41242282/)).
- **Cognition:** Current studies measure brain activity or attention during single walks ([Kupchenko et al., 2023](https://pubmed.ncbi.nlm.nih.gov/36940180/); [Shapiro et al., 2022](https://pubmed.ncbi.nlm.nih.gov/36247999/)); no trial has tested whether months of backward-walking training change cognition, which could either support or refute the attention-training idea.

  
## Conclusion

Backward walking is walking in reverse, on open ground or a treadmill. It changes how the thigh and calf muscles work, may ease pressure behind the kneecap (shown so far only for backward running), takes more effort at a slow pace, and challenges balance because the path behind cannot be seen.

The firmest support comes from people with knee arthritis, where adding backward walking to physical therapy has repeatedly eased pain and improved daily function, and from people relearning to walk after stroke, where balance has improved modestly; results in other nerve conditions are inconsistent. Stronger legs are supported by smaller studies. Claims about fitness, weight, inflammation, back pain, thinking skills, and fall prevention in healthy older adults rest on thin or early evidence.

Whether walking backward does more than the same amount of ordinary walking remains unsettled, and several direct comparisons found little difference. The main risks are falls while learning, especially for frail or unsteady people, and short-lived muscle soreness after downhill or long sessions; supervised studies have reported few serious problems.

The evidence base consists mostly of small, short studies with little long-term follow-up. Most come from physical therapy researchers, whose profession delivers the method, and one prominent promoter sells backward-walking equipment. For health-focused adults who already exercise and value variety, backward walking is a low-cost option whose specific advantage over regular walking is still uncertain.

**[Top](#top) - [Benefits](#expected-benefits) - [Risks](#potential-risks--side-effects) - [Protocol](#therapeutic-protocol)**