Validity of trunk extensor and flexor torque measurements using isokinetic dynamometry

The researchers tested whether a specific isokinetic dynamometer (a machine that measures muscle force at a controlled speed) could accurately and consistently measure trunk flexor (abdominal) and extensor (lower back) muscle strength. They compared three types of muscle contractions:

**Eccentric** (muscle lengthening while under tension — like lowering a weight slowly)

**Isometric** (muscle contracting without movement — like a plank hold)

**Concentric** (muscle shortening while under tension — like a sit-up)

They then validated these torque measurements against two gold-standard biological measures:

1. **Muscle cross-sectional area (CSA)** — measured via MRI, which directly reflects muscle size

2. **Surface electromyography (EMG)** — electrical activity recorded from electrodes on the skin over the muscles, which reflects how hard the muscles are working

The outcome measures were peak torque (maximum force produced) for each contraction type, and the relationship between torque and both CSA and EMG activity.

**15 healthy subjects** (no specific age range given, but typical for this lab: likely 20–40 years old)

Both sexes included (exact ratio not specified, but typical for such studies is roughly equal)

No history of low back pain, neuromuscular disorders, or recent trunk injuries

Recreationally active but not elite athletes

Setting: a university-based exercise physiology laboratory in France

This is a very small sample, which limits how much you can generalise to other populations (e.g., older adults, people with back pain, athletes).

**Isokinetic dynamometer:** A Con-Trex® MJ (CMV AG, Switzerland) trunk module — a specialised chair that locks the pelvis and thighs, with a padded chest harness attached to a motorised arm that moves at a constant speed (60°/s for concentric/eccentric, and a fixed position for isometric)

**Torque measurement:** Peak torque in Newton-metres (Nm) — the rotational force produced by the trunk muscles

**Muscle CSA:** Magnetic resonance imaging (MRI) of the trunk at the L3-L4 vertebral level, measuring the cross-sectional area of the erector spinae (back) and rectus abdominis (abdominal) muscles in square centimetres (cm²)

**EMG:** Bipolar surface electrodes placed over the erector spinae (3 cm lateral to L3 spinous process) and rectus abdominis (3 cm lateral to the umbilicus), recording electrical activity in microvolts (µV) during submaximal isometric contractions at 20%, 40%, 60%, and 80% of maximum voluntary contraction

**Reliability:** Test-retest design with two identical sessions separated by 7 days

**Study design:** Observational, test-retest reliability and validity study. This is not a randomised controlled trial — there is no intervention, no control group, and no blinding. Instead, the researchers are establishing whether a measurement tool (the dynamometer) produces numbers that correspond to biological reality (validity) and are stable over time (reliability).

**Procedure:** Each subject attended two sessions, 7 days apart, at the same time of day. In each session:

1. Standardised warm-up (5 minutes of submaximal trunk flexion/extension on the dynamometer)

2. Familiarisation trials (3–5 submaximal contractions for each contraction type)

3. Maximal voluntary contractions (3 trials each for concentric, eccentric, and isometric trunk flexion and extension, with 60-second rest between trials)

4. Submaximal isometric contractions at 20%, 40%, 60%, and 80% of maximum (for EMG validation)

**Why this design matters:**

**Test-retest reliability** (same test on two separate days) tells you whether a single measurement is trustworthy. If you test your trunk strength today and again next week, a reliable machine should give you nearly the same number. The 7-day gap is long enough to avoid muscle fatigue or learning effects, but short enough that true strength changes are unlikely.

**Validity** is established by correlating torque with muscle CSA (structural validity) and EMG (construct validity). If torque truly reflects muscle strength, it should correlate with how much muscle you have (CSA) and how electrically active that muscle is during contraction (EMG).

**No randomisation or blinding** is needed here because there's no treatment being tested. The design is purely about measurement properties.

**What this design can prove:**

That the dynamometer produces numbers that are consistent (reliable) and biologically meaningful (valid) for healthy young adults

That the measurement error is small enough to detect real changes in strength (e.g., from training or injury)

**What this design cannot prove:**

Whether the dynamometer is valid for people with back pain, older adults, or athletes

Whether trunk strength measured this way predicts real-world outcomes (like lifting ability or injury risk)

Whether the machine is better or worse than other trunk strength testing methods (no comparator device was tested)

**Major methodological weaknesses:**

Very small sample (n=15) — reliability estimates from small samples can be unstable

No blinding of testers or subjects (not critical for this design, but could introduce subtle bias in encouragement or positioning)

Only healthy subjects — validity may differ in pathological populations

Only one dynamometer model tested — results may not generalise to other brands

**Validity (correlation with muscle size):**

Muscle CSA correlated strongly with peak torque for all contraction types:

- Erector spinae (back extensors): r = 0.74–0.85 (P < 0.001 for all)

- Rectus abdominis (abdominal flexors): r = 0.74–0.85 (P < 0.001 for all)

This means muscle size explains roughly 55–72% of the variation in torque (r² = 0.55–0.72)

**Validity (correlation with muscle activation):**

EMG activity correlated almost perfectly with submaximal isometric torque:

- Erector spinae: r ≥ 0.99 (P < 0.05)

- Rectus abdominis: r ≥ 0.99 (P < 0.05)

This means the dynamometer's torque readings directly reflect how hard the muscles are actually contracting

**Test-retest reliability:**

Intraclass correlation coefficients (ICC): 0.87–0.95 across all contraction modes

- ICC > 0.90 is considered "excellent" reliability

Standard error of measurement (SEM): less than 9% for all contraction types

- This means if your true strength is 100 Nm, a single test might give you a value between 91 and 109 Nm due to measurement error alone

Mean difference between test and retest: -3.7% to +3.7% (no significant directional bias)

- This means the machine doesn't systematically over- or under-estimate on the second test

**No significant differences** were found between test and retest for any contraction mode (all P > 0.05), confirming no learning effect or fatigue bias.

The correlation between muscle size and strength (r ≈ 0.80) means that if you have 10% more trunk muscle CSA than someone else, you'd expect roughly 8% higher peak torque — but individual variation is large (about 30% of strength differences are not explained by muscle size alone)

The test-retest error of ~9% means that if you test your trunk strength today and again next week, a change of less than 9% could simply be measurement noise. To be confident you've actually gotten stronger, you'd need to see a change greater than about 18% (2 × SEM, the minimal detectable change)

The EMG-torque correlation of r ≥ 0.99 is essentially perfect — the dynamometer captures the electrical command your brain sends to your muscles with almost no distortion

**Acknowledged by authors:**

Small sample size (n=15) limits generalisability

Only healthy, young subjects — validity in patients with back pain or older adults is unknown

Only one dynamometer model tested

EMG was only measured during isometric contractions, not concentric or eccentric

**Additional critical notes:**

**No blinding:** Testers knew which session was which, which could subtly influence how they positioned subjects or provided encouragement

**Familiarisation:** Subjects had only 3–5 submaximal trials before maximal testing. For a self-experimenter, you'd want more practice sessions to ensure true maximal effort

**Population limits:** 15 healthy subjects (likely university students/staff) means results may not apply to:

- Older adults (muscle quality changes with age)

- People with chronic low back pain (pain inhibition alters voluntary activation)

- Elite athletes (different muscle architecture and neural drive)

**No female-specific analysis:** Sex differences in trunk strength are well-documented (men are ~30–50% stronger), but the study didn't report separate reliability estimates for men vs. women

**Single joint angle for isometric:** Isometric testing was done at only one trunk angle (neutral, 0° flexion). Strength at other angles might show different reliability

**Industry funding:** The dynamometer manufacturer (CMV AG) may have provided equipment or support, though the authors declared no conflicts of interest

For someone running their own n=1 experiment on trunk strength:

**What to test:**

Specific intervention: Any trunk-strengthening program (e.g., deadlifts, planks, back extensions, rotational exercises)

Dose: Aim for 3–4 sets of 8–12 repetitions at 60–80% of your 1-rep max (or RPE 7–8), 2–3 times per week

Comparator: Either a control period (no training for 4 weeks) or a different exercise modality (e.g., isometric planks vs. dynamic back extensions)

**Minimum meaningful duration:**

At least 8 weeks of consistent training (2–3 sessions/week)

Why: True strength gains from neural adaptation take 4–6 weeks; muscle hypertrophy takes 8–12 weeks. Shorter periods risk confusing measurement error with real change

Test at baseline, week 4, and week 8 (at minimum)

**What to measure:**

Primary metric: Peak torque (Nm) for trunk extension and flexion, tested at 60°/s (concentric and eccentric) and at neutral position (isometric)

Secondary metrics:

- Rate of torque development (how fast you can produce force — relevant for athletic performance)

- Bilateral symmetry (left vs. right side differences >10% may indicate imbalance)

- Subjective effort (RPE or perceived exertion during testing)

Use the same dynamometer, same time of day, same warm-up protocol, and same tester for all sessions

**Key confounds to control for:**

**Learning effect:** The first test session is almost always lower than subsequent ones. Do at least 2 familiarisation sessions before collecting baseline data

**Time of day:** Trunk strength varies by ~5–10% across the day (typically highest in late afternoon). Test at the same time (±1 hour) for all sessions

**Prior activity:** No heavy lifting or trunk exercise 48 hours before testing. Standardise warm-up (5 minutes light cardio + 3 submaximal practice trials)

**Motivation:** Maximal strength tests require genuine effort. Use standardised verbal encouragement ("Push! Push! Push!") and consider using a visual biofeedback display

**Positioning:** Small changes in hip angle or chest harness tightness can alter torque by 10–15%. Mark positions on the dynamometer with tape and photograph for replication

**Fatigue:** Leave 60–90 seconds rest between maximal trials. If you feel your performance dropping, stop and rest longer

**Menstrual cycle (if applicable):** Strength may vary by 5–10% across the cycle (highest in follicular phase). Test during the same phase for all sessions

**What a positive result would look like:**

A change in peak torque greater than 18% (the minimal detectable change from this study) for any contraction mode

Example: If your baseline concentric trunk extension is 200 Nm, a follow-up value of 236 Nm or higher would be a real increase (not measurement noise)

Consistent improvement across multiple contraction types (e.g., both concentric and eccentric torque increase) strengthens confidence

If you're comparing two interventions (e.g., deadlifts vs. back extensions), a difference of >18% between conditions would be meaningful

For submaximal measures (like EMG or rate of torque development), smaller changes may be detectable — but this study didn't provide minimal detectable change values for those metrics

**Bottom line for your self-experiment:** This dynamometer is trustworthy enough to detect real strength changes, but you need to account for the ~9% measurement error. Plan for at least 8 weeks of training, use familiarisation sessions, control for time of day and prior activity, and look for changes >18% to be confident you've actually gotten stronger. If you're testing at home without a dynamometer, these principles still apply — just use a different strength metric (e.g., 1-rep max on a deadlift or plank hold time) and establish your own test-retest reliability first.