SECTION 11How To Use This — And Where Your Job Ends
Scope of practice — read this once, properly
You are not a physician. You cannot order these tests, you cannot diagnose from them, and you cannot tell a client to start or stop a medication. Nothing in this file changes that, and no amount of understanding substitutes for a licence.
What this file is for: a client hands you a report and asks what it means. You need to read it intelligently, know which lines your training will actually move, know which ones warrant "please ask your doctor about this specific value," and know when a number is alarming enough that the session does not happen today.
The safe sentence, and you should use it often: "Here is what that marker generally reflects, and here is why I think it is worth asking your doctor about. I am not able to interpret it for you."
Referring early costs you nothing. Referring late can cost someone a great deal.
Reference range is not the same as optimal range
A laboratory reference range is the middle 95% of the population that lab tested — not a health target. If the sampled population is largely sedentary and metabolically unwell, "normal" encodes that. This is exactly what happened with ALT: standard upper limits of 40–56 U/L were derived from cohorts containing undiagnosed fatty liver, and when Prati and colleagues re-derived them from genuinely healthy donors in 2002 they landed at 30 U/L for men and 19 for women.
Throughout this file the "Lab normal" column is what the report will flag, and "Optimal" is where the outcome evidence points. A result inside the reference range is not automatically a good result.
Evidence grading used in every table
| Grade | Meaning | Example |
|---|
| A | Large outcome trials or Mendelian randomisation. Causal link to disease is well established. | ApoB, blood pressure, HbA1c, Lp(a) |
| B | Strong, consistent observational association. Causality plausible but not proven. | hs-CRP, homocysteine, ferritin |
| C | Useful clinically but indirect, or heavily confounded. | Free T3, reverse T3, DHEA-S |
| D | Weak, inconsistent, or mostly marketing. | Telomere length, most food-sensitivity panels |
Grades are my summary of the strength of the causal evidence, not of how commonly a test is sold. Several heavily marketed panels sit at grade D. Being honest about this with clients is a differentiator — most of their other sources will not be.
The uncomfortable hierarchy
Ranked by strength of the mortality evidence behind them: 1. VO₂max — the single most powerful predictor of all-cause mortality available, with effect sizes exceeding those of smoking cessation at the low end. 2. Grip strength — remarkably predictive and measurable with a cheap dynamometer. 3. Blood pressure. 4. ApoB and Lp(a). 5. HbA1c and fasting insulin. Then a long gap, and only then the fashionable longevity panels.
Two of the top three are things you measure and improve directly. That is the argument for your profession, and it is a strong one.
The over-testing trap
More markers means more false positives. A reference range covering 95% of people means 1 in 20 healthy results falls outside it by chance alone — so a client tested on 60 markers should expect roughly three abnormal values while being entirely well. Each one generates anxiety, follow-up testing, and sometimes an unnecessary procedure.
The correct advice is rarely "test everything." It is: test the things that would change what you do.
SECTION 12Tier 1 — The Foundational Panel
What almost any client will already have from a routine health check. If they bring you one report, it is this one.
| Marker | Lab "normal" | Optimal | What it actually measures | Why it predicts outcome | What moves it | Refer when | Ev |
|---|
| ── GLYCAEMIC CONTROL — the metabolic core ── |
| Fasting glucose | 70–99 mg/dL | 75–90 | Glucose after 8–12 h fast | Late marker — rises only after beta cells are already failing. Normal fasting glucose does not rule out insulin resistance. | Diet, weight loss, aerobic + resistance training | ≥100 (pre-diabetic) · ≥126 twice (diabetic) | A |
| HbA1c | <5.7% | <5.4% | Glycated haemoglobin — 2–3 month average glucose | Each 1% rise carries a substantial increase in cardiovascular and microvascular risk. The workhorse of metabolic assessment. | Diet, training, weight loss. Falls over 8–12 weeks. | 5.7–6.4 pre-diabetes · ≥6.5 diabetes | A |
| Fasting insulin | 2–25 µIU/mL | <8, ideally 2–5 | How hard the pancreas is working to hold glucose down | Rises years before glucose does. The earliest routinely available signal of metabolic dysfunction — and rarely ordered. | Very responsive to training and fat loss | >12 with normal glucose = compensated insulin resistance | B |
| HOMA-IR | <2.5 | <1.5 (Asian cutoffs run lower) | Calculated: (glucose mg/dL × insulin) ÷ 405 | Combines the two above into one insulin-resistance index. Asian populations show dysglycaemia at lower values (~1.4–2.5). | Training, fat loss, sleep | >2.5, or >2.0 in an Indian client | B |
| ── LIPIDS — standard panel ── |
| Total cholesterol | <200 mg/dL | Not useful alone | All cholesterol in all particles | Nearly useless in isolation — high HDL inflates it, and it says nothing about particle count. Ignore it in favour of the rows below. | Modestly by diet | Only in context of the full panel | C |
| LDL-C | <100 mg/dL | <70 if any other risk | Cholesterol carried inside LDL particles | Causal for atherosclerosis, but it measures cargo not vehicles. Can read normal while particle count is high — see ApoB. | Saturated fat, fibre, weight loss, statins. Exercise alone moves it little. | ≥160 · ≥190 = likely familial, urgent | A |
| HDL-C | >40 ♂ / >50 ♀ | 50–70 | Cholesterol in HDL particles | Low HDL predicts risk, but raising it pharmacologically has never reduced events. A marker, not a target. | Aerobic exercise, weight loss, quitting smoking | <40 ♂ / <50 ♀ as part of metabolic syndrome | B |
| Triglycerides | <150 mg/dL | <90 | Circulating fat, mostly in VLDL | One of the most training-responsive lipids you have. A high TG with low HDL is the classic insulin-resistance signature. | Very responsive — alcohol, refined carbs, training, fat loss | ≥200 · ≥500 = pancreatitis risk, urgent | B |
| Non-HDL cholesterol | <130 mg/dL | <100 | Total minus HDL — all atherogenic particles | Better than LDL-C, free to calculate, no extra test. A decent stand-in when ApoB is unavailable. | As per LDL and triglycerides | ≥160 | A |
| ── LIVER ── |
| ALT | <40–56 U/L | <30 ♂ / <19 ♀ | Liver cell enzyme; leaks when hepatocytes are stressed | The standard lab range is too permissive — it was set on populations containing undiagnosed fatty liver. Prati 2002 derived true limits of 30 and 19. | Fat loss and training reverse fatty liver, often dramatically | >2× upper limit, or persistently raised | B |
| AST | <40 U/L | <30 | Also found in muscle and heart | Rises after hard training — muscle damage, not liver. AST:ALT above 2 suggests alcohol. | Confounded by recent exercise — test 48 h after a hard session | Persistent elevation without exercise explanation | B |
| GGT | <55 U/L | <25 | Biliary enzyme; sensitive to alcohol and oxidative stress | Independently predicts cardiovascular disease and mortality even within the "normal" band. An underrated marker. | Alcohol reduction, fat loss | Raised with ALP, or a rising trend | B |
| Albumin | 3.5–5.0 g/dL | 4.2–5.0 | Main plasma protein; synthetic liver function | Low albumin predicts mortality across nearly every disease state. Reflects nutrition, inflammation and liver capacity together. | Protein intake, resolving inflammation | <3.5 | B |
| ── KIDNEY ── |
| Creatinine | 0.7–1.3 mg/dL ♂ | Interpret with muscle mass | Muscle metabolism waste product | Rises with muscle mass and creatine supplementation. A muscular client can look like they have kidney disease and not have it. | Muscle mass, creatine, protein, hydration | Rising trend — but check creatine use first | B |
| eGFR (creatinine-based) | >90 mL/min/1.73m² | >90 | Estimated filtration rate | Assumes creatinine production is constant — an assumption that breaks in muscular and creatine-supplementing clients. | Not directly trainable | <60 on two occasions three months apart | A |
| Cystatin C | 0.5–1.0 mg/L | <0.9 | Filtration marker made by all nucleated cells | Independent of muscle mass, diet and creatine. The test to request when creatinine looks falsely alarming. | Not trainable | Use to confirm or refute a low creatinine-based eGFR | A |
| Uric acid | 3.5–7.2 mg/dL ♂ | <6.0 | Purine metabolism end product | Associated with hypertension, metabolic syndrome and gout. Fructose and alcohol raise it. | Fructose, alcohol, purine load, weight loss | >7 with joint pain | B |
| Urine albumin:creatinine | <30 mg/g | <10 | Protein leaking into urine | Detects kidney damage far earlier than eGFR, and independently predicts cardiovascular events. Cheap and underused. | Blood pressure and glucose control | >30 on two samples | A |
| ── COMPLETE BLOOD COUNT ── |
| Haemoglobin | 13–17 g/dL ♂ / 12–15 ♀ | Mid-range | Oxygen-carrying capacity | Low Hb directly limits VO₂max and training capacity. High-normal in a smoker or apnoea patient can signal chronic hypoxia. | Iron, B12, folate status; altitude | Below range, or a falling trend | A |
| MCV | 80–100 fL | 85–92 | Average red cell size | Small cells point to iron deficiency; large cells to B12 or folate deficiency, or alcohol. Directs the next test. | Correcting the underlying deficiency | Outside range — pairs with ferritin and B12 | A |
| RDW | 11.5–14.5% | <13 | Variation in red cell size | A raised RDW predicts all-cause mortality across many conditions — one of the most quietly powerful numbers on a routine CBC. | Nutritional correction | >15 warrants investigation | B |
| Neutrophil:lymphocyte ratio | 1–3 | <2 | Calculated from the differential — free | A simple systemic inflammation index. Elevated NLR is associated with cardiovascular events and poorer outcomes generally. | Resolving infection, reducing chronic inflammation | >3 persistently | B |
| Platelets | 150–400 ×10⁹/L | Mid-range | Clotting cells | Low platelets raise bleeding risk — relevant before contact or heavy training. | Not trainable | Outside range | A |
Fasting insulin is the one to push for. It is cheap, rarely included by default, and it rises years before fasting glucose does. A client with perfect glucose and an insulin of 18 is not metabolically healthy — they are compensating, and the compensation is running out. Paired with glucose it also gives you HOMA-IR for free.
Two liver traps to avoid. First, the standard ALT range is too permissive — use 30 (men) and 19 (women) as your mental threshold. Second, and more likely to bite you: AST and ALT rise after hard training, because AST in particular is abundant in skeletal muscle. A client who squatted heavy two days before a blood draw can produce a liver panel that looks alarming and means nothing. Have them rest 48–72 hours before testing, and if a report looks bad, ask what they trained first.
SECTION 13Tier 2 — The Markers That Change Risk Assessment
These are not usually on a routine panel, and they are where the useful information lives. If a client is going to pay for extra tests, these are the ones worth their money.
| Marker | Lab "normal" | Optimal | What it actually measures | Why it predicts outcome | What moves it | Refer when | Ev |
|---|
| ── ADVANCED LIPIDS — where the real signal is ── |
| ApoB | <130 mg/dL | <80 general · <70 higher risk · <60 very high | One ApoB molecule per atherogenic particle — so it counts the particles, not the cargo | Every LDL, VLDL, IDL and Lp(a) particle carries exactly one ApoB. It therefore counts the vehicles that can lodge in an artery wall. Where ApoB and LDL-C disagree, ApoB tracks risk better. National Lipid Association 2024 targets: <90 intermediate, <70 high, <60 very high risk. | Saturated fat, fibre, weight loss, statins. Exercise helps modestly. | >100, or discordant with a "normal" LDL-C | A |
| Lp(a) | <30 mg/dL | <30 mg/dL (<75 nmol/L) | A genetically determined, particularly atherogenic LDL variant | The most important single test for your client base. Roughly 25% of South Asians exceed 50 mg/dL, versus far fewer in European populations — a major contributor to why South Asians have heart attacks around a decade earlier. Causal by Mendelian randomisation. Measure once in a lifetime — it is ~90% genetic and barely moves. | Almost nothing. Not diet, not training. Newer drugs are in trials. | >50 mg/dL (>125 nmol/L) — and screen first-degree relatives | A |
| Remnant cholesterol | No standard | <20 mg/dL | Calculated: total − HDL-C − LDL-C | Triglyceride-rich remnant particles are causally atherogenic. Free to calculate from any standard panel. | Highly responsive to carbohydrate, alcohol and fat loss | >30 | B |
| LDL particle number (LDL-P) | <1000 nmol/L | <1000 | Direct particle count by NMR | Conceptually similar to ApoB. ApoB is cheaper and better standardised — prefer it unless LDL-P is what the client already has. | As per ApoB | >1300 | B |
| ── INFLAMMATION — the practical answer to "cytokines" ── |
| hs-CRP | <3 mg/L | <1 mg/L | Acute-phase protein made by the liver in response to IL-6 | The inflammation marker that actually earns its place. AHA/CDC bands: <1 low risk, 1–3 average, >3 high. JUPITER showed people with normal LDL but raised hs-CRP still carried elevated risk. Cheap, standardised, large outcome datasets. | Fat loss (especially visceral), training, sleep, smoking cessation | >3 persistently · >10 means acute illness — retest in 2 weeks | A |
| IL-6, TNF-α | Assay-dependent | — | The upstream cytokines that drive CRP production | Research tools, not clinical ones. Expensive, poorly standardised between labs, and swing hour to hour. Nothing you would do differently based on the result. If a client asks about "testing cytokines," hs-CRP is the answer. | Same as hs-CRP | Not routinely useful — do not recommend | D |
| ESR | <15 ♂ / <20 ♀ mm/hr | <10 | How fast red cells settle — an indirect inflammation measure | Slower and less specific than CRP. Still used for rheumatological conditions. | Resolving underlying inflammation | Markedly raised, especially with joint symptoms | B |
| Fibrinogen | 200–400 mg/dL | <300 | Clotting protein and acute-phase reactant | Links inflammation to thrombotic risk. Independently associated with cardiovascular events. | Training, smoking cessation, fat loss | >400 persistently | B |
| Homocysteine | <15 µmol/L | <9 | Amino acid intermediate that accumulates when B12, folate or B6 are low | Strongly associated with vascular disease and cognitive decline. But lowering it with B vitamins has not reduced events in trials — it may be a marker rather than a cause. Still useful as a functional B-vitamin readout. | B12, folate, B6 repletion | >12 — check B12 and folate | B |
| ── IRON & MICRONUTRIENT STATUS ── |
| Ferritin | 30–300 ng/mL ♂ / 15–200 ♀ | 50–150 | Stored iron — and an acute-phase reactant | Dual-natured and often misread. Low ferritin means iron deficiency even with normal haemoglobin, and causes fatigue and poor training tolerance. But inflammation falsely raises it — always interpret alongside hs-CRP. | Iron intake, vitamin C, blood loss. Endurance athletes run low. | <30 (or <50 with symptoms) · >300 needs investigation | A |
| Transferrin saturation | 20–50% | 25–40% | Proportion of iron transport capacity in use | Cuts through the ferritin ambiguity. Low saturation with high ferritin means inflammation, not iron overload. | Iron status | <20% (deficiency) · >45% (possible haemochromatosis) | A |
| Vitamin D (25-OH) | >20 ng/mL | 30–50 ng/mL | Circulating vitamin D storage form | Deficiency is close to a default assumption in urban India. Associated with muscle function, bone density and mood. Note that supplementation trials have been largely disappointing for hard outcomes — correct a deficiency, do not chase a high number. | Sunlight, supplementation, magnesium status | <20 deficient · <30 insufficient · >100 toxic | B |
| Vitamin B12 | >200 pg/mL | >400 pg/mL | Circulating B12 — much of it not bioavailable | The reference range is too low. Neurological damage occurs in the 200–400 "normal" band. Long-term vegetarians and metformin users will be low — extremely relevant in India. | Supplementation — animal foods only otherwise | <400 with symptoms — confirm with MMA and homocysteine | A |
| Methylmalonic acid (MMA) | <0.4 µmol/L | <0.27 | Accumulates when B12 is functionally deficient | More reliable than serum B12 itself. Order when B12 sits in the ambiguous 200–400 range. | B12 repletion | Raised = true B12 deficiency regardless of serum B12 | A |
| RBC magnesium | 4.2–6.8 mg/dL | Upper half | Intracellular magnesium | Serum magnesium is nearly useless — the body defends it tightly at the expense of tissue stores. RBC magnesium is the better test. | Diet, supplementation (citrate or glycinate) | Low with cramping or arrhythmia | C |
| Omega-3 index | No standard | >8% | EPA + DHA as a percentage of red cell fatty acids | Below 4% is associated with higher cardiovascular mortality in observational data. Most Indian vegetarian diets sit low. | Fatty fish, algal oil or fish oil supplementation | <4% | B |
Lp(a) deserves its own paragraph, because of who your clients are. Roughly 25% of South Asians have Lp(a) above 50 mg/dL, a far higher proportion than in European populations, and it is a leading candidate explanation for why South Asians suffer myocardial infarction on average about a decade earlier and with more aggressive, multi-vessel disease.
It is around 90% genetically determined, so it needs measuring once in a lifetime — and almost nothing you or they do will change it. That sounds discouraging, but it is not. A high Lp(a) is a reason to be considerably more aggressive about every risk factor that is modifiable, and a reason to suggest first-degree relatives get tested too. It is arguably the single highest-value test on this page for an Indian client base.
On your cytokine question. Your instinct was right that inflammation matters; the practical answer is just different from what you would expect. IL-6 and TNF-α are genuine inflammatory cytokines, but they are research tools — expensive, poorly standardised between laboratories, and variable hour to hour. Nothing you would do changes based on the result.
hs-CRP is the one that earns its place. It is downstream of IL-6, cheap, well standardised, and carries large outcome datasets. Under 1 mg/L is low risk, over 3 is high. Add fibrinogen and the neutrophil-to-lymphocyte ratio (free, calculated from any CBC) if you want more depth. Above 10 mg/L means acute illness — retest in a fortnight rather than drawing conclusions.
SECTION 14Tier 3 — Hormones
The area where trainers most often overstep. Every row below is written to help you recognise a pattern and refer, not to interpret or advise. Hormone results are also the most sensitive to when and how the sample was taken — see section 16 before you take any of them at face value.
| Marker | Lab "normal" | Optimal | What it tells you | Why it matters | What moves it | Refer when | Ev |
|---|
| ── THYROID ── |
| TSH | 0.4–4.5 mIU/L | 0.5–2.5 | Pituitary signal — rises when thyroid output is inadequate | Inverse to thyroid function: high TSH means an underactive gland. Sets metabolic rate, and therefore fatigue, weight and training tolerance. Biotin supplements falsely distort this test. | Not trainable. Iodine and selenium support the gland. | >4.5 or <0.4 — needs free T4 | A |
| Free T4 | 0.8–1.8 ng/dL | Mid-to-upper half | The circulating thyroid prohormone | Interpreted with TSH. High TSH plus low free T4 is overt hypothyroidism; high TSH with normal free T4 is subclinical. | Not trainable | Outside range | A |
| Free T3 | 2.3–4.2 pg/mL | Mid-to-upper half | The active hormone | Falls in a prolonged calorie deficit — a genuine adaptive response, not thyroid disease. Useful context for a stalled fat-loss client. | Energy availability. Rises when calories are restored. | Low with symptoms — but consider under-eating first | C |
| TPO antibodies | <35 IU/mL | Negative | Autoimmune attack on the thyroid | Positive TPO indicates Hashimoto’s, the commonest cause of hypothyroidism. Explains a rising TSH before it leaves the range. | Not trainable | Positive — endocrinology follow-up | A |
| ── MALE HORMONES ── |
| Total testosterone | 264–916 ng/dL | 500–800 | Total circulating testosterone, bound and free | Harmonised reference range for healthy non-obese men 19–39 is 264–916 ng/dL. Endocrine Society uses 264 as the hypogonadism cutoff; AUA uses 300. Diagnose only with symptoms plus two low morning samples. | Fat loss, sleep, resistance training, adequate calories. Chronic under-eating suppresses it. | <300 on two morning samples with symptoms | A |
| Free testosterone | 5–21 ng/dL | Upper half | The unbound, biologically active fraction | More informative than total when SHBG is abnormal. Obesity lowers SHBG and can mask a low total reading. | As per total testosterone | Low despite normal total | B |
| SHBG | 10–57 nmol/L | 20–45 | Binding protein that controls how much testosterone is free | Low SHBG accompanies insulin resistance and obesity; high SHBG accompanies under-eating and hyperthyroidism. Needed to interpret total testosterone properly. | Insulin sensitivity, body fat, energy availability | Markedly abnormal | B |
| Estradiol (men) | 10–40 pg/mL | 20–30 | Testosterone aromatised to oestrogen | Both extremes cause problems — men need oestrogen for bone and libido. Higher body fat means more aromatase and more conversion. | Fat loss | >50 with symptoms | B |
| LH and FSH | 1.5–9 IU/L | Mid-range | Pituitary signals to the testes | Separates the cause: low testosterone with high LH is testicular; with low LH it is pituitary or hypothalamic. Suppressed LH with high testosterone means exogenous use. | Energy availability, sleep | Abnormal with low testosterone | A |
| ── FEMALE HORMONES ── |
| Estradiol (women) | Varies by cycle phase | Cycle-appropriate | Primary oestrogen | Timing is everything — a result is uninterpretable without knowing the cycle day. Low estradiol in an athletic woman suggests low energy availability and carries real bone consequences. | Energy availability, body fat, training load | Persistently low, or absent periods for 3+ months | A |
| Progesterone | Peaks mid-luteal | >5 ng/mL mid-luteal | Confirms ovulation occurred | Draw 7 days after ovulation. A low mid-luteal progesterone means an anovulatory cycle — often the first sign of under-fuelling. | Energy availability | Low mid-luteal repeatedly | A |
| LH : FSH ratio | ~1:1 | ~1:1 | Pituitary balance | An LH:FSH ratio above 2:1 is suggestive of PCOS — common, and highly relevant to body composition goals. | Insulin sensitivity, weight loss | >2:1 with irregular cycles | B |
| AMH | 1–4 ng/mL | Age-dependent | Ovarian reserve | Relevant for fertility planning. Raised in PCOS. | Not trainable | Fertility concerns | B |
| ── STRESS & GROWTH AXIS ── |
| Morning cortisol | 6–23 µg/dL | 10–18 at 8 am | Primary stress hormone | A single random draw is nearly meaningless — cortisol follows a steep diurnal curve. Standardise to 8 am, or use a 4-point salivary curve. Commercial "adrenal fatigue" panels are not a recognised diagnosis. | Sleep, training load, psychological stress | Very high or very low — endocrinology | C |
| DHEA-S | 80–560 µg/dL ♂ | Upper half for age | Adrenal androgen precursor | Declines steadily with age. Marketed heavily as an anti-ageing marker; the outcome evidence is thin. | Modestly by training and stress reduction | Markedly low or high | C |
| IGF-1 | Age-dependent | Mid-to-upper for age | Growth hormone axis readout | A genuine tension for longevity: higher IGF-1 supports muscle and bone, but lower IGF-1 associates with longer lifespan in animal models. No consensus on an optimal target in humans. | Protein intake, sleep, training. Falls with fasting. | Very high (rule out acromegaly) or very low | C |
Three rules that will keep you out of trouble here.
1. Timing invalidates more hormone results than pathology does. Testosterone drawn at 4 pm can read 20–30% below the same person at 8 am. A female hormone panel without a recorded cycle day cannot be interpreted at all. Cortisol without a time stamp is meaningless.
2. Low hormones in a lean, hard-training client usually mean under-fuelling, not disease. Suppressed free T3, low testosterone, high SHBG, absent periods — this is the low energy availability picture, and the fix is food, not an endocrinologist. Though if periods have been absent three months or more, it needs both.
3. Never comment on whether someone should take a hormone. Not testosterone, not thyroid medication, not anything. Recognise the pattern, describe what you have noticed, refer. That is the whole of your role.
SECTION 15Sleep & HRV
Sleep is the highest-yield recovery intervention available and it costs nothing. Undiagnosed sleep apnoea will quietly undermine every training block you write — and in a client who is overweight, snores and is always tired, it is common and routinely missed.
STOP-BANG — screening for obstructive sleep apnoea
| Letter | Question | Scores a point if |
|---|
| S — Snoring | "Do you snore loudly? Loud enough to be heard through a closed door?" | Yes |
| T — Tired | "Do you often feel tired, fatigued or sleepy during the daytime?" | Yes |
| O — Observed | "Has anyone observed you stop breathing, choke or gasp during sleep?" | Yes |
| P — Pressure | "Do you have, or are you treated for, high blood pressure?" | Yes |
| B — BMI | Body mass index | >35 kg/m² (use a lower threshold for Indian clients) |
| A — Age | Age in years | Over 50 |
| N — Neck | Neck circumference | >43 cm men / >41 cm women |
| G — Gender | Sex | Male |
| Score | Risk of obstructive sleep apnoea | What you do |
|---|
| 0 – 2 | Low | Moderate-to-severe OSA can be confidently ruled out. No action needed. |
| 3 – 4 | Intermediate | Mention it. A score of ≥3 catches 93% of moderate-to-severe cases — sensitive, but many false positives. |
| 5 – 8 | High | Refer for a sleep study. A score ≥5 identifies moderate-to-severe OSA with 94% specificity. |
This takes ninety seconds and you can do it in an intake conversation. A score of ≥3 is highly sensitive — it catches 93% of moderate-to-severe cases — but produces plenty of false positives, so treat it as a prompt to mention, not a finding. A score of ≥5 is far more specific and clearly warrants referral for a sleep study.
Note the BMI criterion. The original threshold is 35, but given that Indian clients develop metabolic disease at markedly lower BMI, treat a lower figure as meaningful in your population rather than waiting for 35.
Apnoea-hypopnoea index — what a sleep study reports back
| AHI (events per hour) | Severity | What it means for your client |
|---|
| <5 | Normal | No sleep-disordered breathing. |
| 5 – 15 | Mild | Daytime sleepiness likely. Weight loss often improves it meaningfully. |
| 15 – 30 | Moderate | Meaningful cardiovascular and metabolic consequences. Treatment indicated. |
| >30 | Severe | Substantially raised cardiovascular risk. Under-recovery will sabotage every training block until treated. |
Sleep metrics worth tracking
| Metric | Target | Why it matters to training | Note |
|---|
| Total sleep time | 7–9 hours | Below 6 h impairs glucose tolerance, raises appetite, lowers testosterone and cuts time to exhaustion. | The highest-yield recovery intervention you can prescribe, and it is free. |
| Sleep efficiency | >85% | Time asleep ÷ time in bed. Low efficiency with adequate time in bed suggests insomnia or apnoea. | Wearables estimate this reasonably well. |
| Sleep latency | 10–20 min | Under 5 minutes consistently indicates significant sleep debt. | Falling asleep "instantly" is not a sign of good sleep. |
| Resting heart rate | Individual baseline | A rise of 5+ bpm over 3 mornings is a reliable early overreaching signal. | Cheaper and more robust than HRV for this purpose. |
| Overnight SpO₂ dip | No dips below 90% | Repeated desaturation suggests apnoea. | Consumer wearables are unreliable here — use as a prompt to refer, not a diagnosis. |
Heart rate variability — useful, and widely over-read
| Age band | Typical RMSSD (ms) | What a good trend looks like | What a bad trend looks like |
|---|
| 20–29 | 50–70 | Stable or slowly rising baseline | Multi-week decline with rising resting HR |
| 30–39 | 40–60 | Recovers within a day after a hard session | Stays suppressed for 3+ days after normal training |
| 40–49 | 30–50 | Predictable weekly rhythm | Erratic, large day-to-day swings |
| 50–59 | 25–40 | Rises across a training block | Falls across a block despite deloads |
| 60+ | 15–30 | Maintained year on year | Paradoxically high with fatigue — see the note below |
What HRV is good for: tracking your own client against their own baseline over weeks. A sustained multi-week decline alongside a rising resting heart rate and subjective fatigue is a genuine overreaching signal.
What it is not good for: comparing one person to another. Two people the same age can differ by 30–40 ms for entirely benign reasons. Day-to-day variation of 10–30% is normal and driven by sleep, alcohol, hydration and stress — so a single low morning reading means very little.
The counterintuitive one: deep overtraining can present with paradoxically high HRV alongside fatigue, low mood and falling performance. High HRV plus poor subjective recovery is a warning sign, not reassurance. And wrist optical sensors do not produce the same absolute values as a chest strap or ECG — never compare a device reading against clinical reference tables.
SECTION 16What Distorts a Test Result
Before you or anyone else reacts to an abnormal number, check this section. A large share of alarming results in fit, supplement-taking clients are artefacts — and spotting one is genuinely valuable to the client.
| What the client took or did | What it does to the result | Which tests | How long before testing | What you say |
|---|
| ── SUPPLEMENTS ── |
| Biotin (B7) | Falsely low or falsely high depending on assay design — it competes with the biotinylated antibody | TSH, free T4, free T3, testosterone, troponin, hCG, vitamin D | Stop 48–72 h before | "High-dose biotin can make thyroid results look completely wrong in either direction. Stop it two to three days before the draw and tell the lab you were taking it." The FDA issued a safety warning on this in 2017. |
| Creatine | Raises serum creatinine by increasing creatinine production — not by damaging kidneys | Creatinine, creatinine-based eGFR | No need to stop — declare it | "Creatine raises the waste product the kidney test measures, so eGFR reads low even though filtration is perfectly normal. Ask for cystatin C — it is unaffected by muscle mass or creatine." Systematic reviews find no renal harm at 3–5 g/day even beyond a year. |
| High-dose vitamin C | Interferes with some glucose meters and stool occult blood tests | Point-of-care glucose, faecal occult blood | 24 h | Minor, but worth declaring if a result looks odd. |
| Iron supplements | Transiently raises serum iron and transferrin saturation | Iron studies | Stop 24–48 h before | "Take the iron test before your morning dose, not after." |
| ── MEDICATIONS ── |
| Statins | Lower LDL-C and ApoB substantially. May raise ALT/AST modestly and CK. | Lipids, liver enzymes, CK | Do not stop — record the dose | A "good" lipid panel on a statin is not the same as a good panel without one. Note the medication next to the result. |
| Metformin | Lowers HbA1c and glucose. Depletes B12 over years. | HbA1c, glucose, B12 | Continue as prescribed | Flag B12 for annual testing in any long-term metformin user. |
| Testosterone (TRT / AAS) | Raises haemoglobin and haematocrit, suppresses LH and FSH, lowers HDL, may raise ALT | CBC, LH, FSH, lipids, liver | Continue — but disclose | Suppressed LH with high testosterone is the signature of exogenous use. Haematocrit above 54% is a genuine thrombotic risk requiring medical management. |
| Corticosteroids | Raise glucose, white cell count and triglycerides. Suppress cortisol axis. | Glucose, HbA1c, CBC, cortisol | Continue — record it | A steroid course can push a client into apparent pre-diabetes temporarily. |
| Levothyroxine | Lowers TSH, raises free T4 | Thyroid panel | Take the test before the morning dose | Taking the tablet first can transiently raise free T4 and confuse dosing. |
| Oral contraceptives | Raise SHBG and total cortisol, alter lipids and thyroid binding | SHBG, free testosterone, lipids | Continue — record it | Total testosterone in a woman on the pill is often uninterpretable without SHBG. |
| ── WHAT THE CLIENT DID BEFOREHAND ── |
| Hard training in the last 48 h | Raises CK, AST, ALT, hs-CRP, white cells, creatinine | Liver enzymes, inflammation, kidney, CK | Rest 48–72 h before the draw | The most common cause of a scary-looking liver result in a gym-going client. A leg day two days before can double AST. |
| Not fasting | Raises triglycerides and glucose substantially | Glucose, insulin, triglycerides, lipid panel | 8–12 h fast, water allowed | ApoB and LDL-C are relatively fast-insensitive. Triglycerides are not. |
| Dehydration | Concentrates everything — haemoglobin, albumin, creatinine all read high | Almost every blood marker | Normal hydration, no fasting-plus-dehydration | A client who fasted 14 h and skipped water will have a subtly wrong panel across the board. |
| Acute illness or infection | Raises hs-CRP, ferritin, white cells; lowers albumin, iron, zinc | Inflammation, iron studies, albumin | Wait 2–4 weeks after recovery | A ferritin of 400 during a chest infection means inflammation, not iron overload. |
| Time of day | Testosterone and cortisol both peak in the morning and fall through the day | Testosterone, cortisol | Draw between 7 and 10 am | An afternoon testosterone draw can read 20–30% lower and produce a false diagnosis. |
| Menstrual cycle phase | Oestrogen and progesterone vary enormously across the cycle | Estradiol, progesterone, LH, FSH | Day 2–4 for baseline; 7 days post-ovulation for progesterone | A female hormone panel without a cycle day recorded is uninterpretable. |
| Alcohol in the last 72 h | Raises GGT, triglycerides, uric acid and MCV | Liver, lipids, uric acid | Avoid 72 h before | A single heavy night can noticeably move GGT and triglycerides. |
The two you will meet most often, and both are in your client population by default.
Creatine and kidney function. Creatine raises serum creatinine because it increases creatinine production from a larger creatine pool — not because it harms the kidney. Every eGFR equation assumes creatinine production is constant, so it misreads the extra creatinine as reduced filtration. Systematic reviews and meta-analyses find no renal impairment in healthy people at 3–5 g/day, even beyond a year. The answer is cystatin C, which is made by all nucleated cells and is unaffected by muscle mass, diet or supplementation. A client told they have early kidney disease on the basis of creatinine alone, who lifts and takes creatine, should have cystatin C before anyone concludes anything.
Biotin and thyroid panels. High-dose biotin is in most hair, skin and nail supplements. It competes with the biotinylated antibodies used in many immunoassays, producing results that are falsely low in some assay designs and falsely high in others — which means it can fake either hypothyroidism or hyperthyroidism. It also distorts troponin, the test used to diagnose heart attacks. The FDA issued a formal safety communication about this in 2017. Stop it 48–72 hours before any blood draw and tell the laboratory.
What your training actually moves — and what it does not
This is where the whole file becomes useful in a conversation. Promise the top block. Be measured about the middle. Be honest about the bottom.
| Marker | Response to training + nutrition | Typical timeframe | Realistic magnitude | How you frame it |
|---|
| ── STRONGLY RESPONSIVE — take credit for these ── |
| Triglycerides | Large | 4–8 weeks | −20 to −50% | "This is the one that moves fastest. Expect a visible change by your next test." |
| Fasting insulin / HOMA-IR | Large | 4–12 weeks | −20 to −40% | "Your pancreas stops having to shout. This changes before the scale does." |
| HbA1c | Large | 8–12 weeks | −0.3 to −1.0% | "It reflects three months, so we test it quarterly, not monthly." |
| hs-CRP | Large | 8–16 weeks | −30 to −50% | "Visceral fat is inflammatory tissue. Lose it and this falls." |
| Blood pressure | Large | 2–8 weeks | −5 to −10 mmHg systolic | "Comparable to what a first-line medication achieves." |
| ALT / GGT (fatty liver) | Large | 8–24 weeks | Can normalise fully | "Fatty liver is one of the most reversible things on your report." |
| VO₂max | Large | 8–12 weeks | +10 to +25% | "The strongest single predictor of how long you live — and we can move it." |
| Visceral fat (DEXA) | Large | 12–24 weeks | −20 to −40% | "The fat that matters medically goes before the fat you can see." |
| ── MODERATELY RESPONSIVE ── |
| HDL-C | Modest | 12–24 weeks | +5 to +10% | "Slow to move, and raising it is not itself the goal." |
| ApoB / LDL-C | Modest from exercise; large from diet | 8–16 weeks | −5 to −15% exercise alone | "Training helps a little. Diet and, if needed, medication do the heavy lifting." |
| Testosterone | Modest, mostly via fat loss and sleep | 12–24 weeks | +10 to +20% if suppressed | "We restore what excess fat and poor sleep took. We do not exceed your ceiling." |
| Bone density | Slow but real | 6–12 months | +1 to +3% per year | "Slow, and worth every session. This is your independence at 75." |
| Resting heart rate | Moderate | 4–12 weeks | −5 to −15 bpm | "A free daily readout that your heart is getting more efficient." |
| ── BARELY OR NOT AT ALL — do not promise these ── |
| Lp(a) | Essentially none | — | ~0% | "This one is genetic. We cannot change it — so we work harder on everything we can." |
| APOE genotype | None | — | 0% | "Fixed at birth. It changes our strategy, not our effort." |
| TSH / thyroid | Essentially none | — | ~0% | "Training does not fix a thyroid. That is a medical conversation." |
| Total cholesterol | Little, and it is the wrong target | — | Variable | "We track ApoB instead. Total cholesterol tells us almost nothing on its own." |
Use this table to sell honestly. Most trainers either over-promise ("we will fix your cholesterol") or say nothing at all. Being the person who can say "your triglycerides and insulin will move a lot and fast, your ApoB will move a little and diet matters more, and your Lp(a) will not move at all — which is exactly why the rest matters so much for you" is a genuinely different level of practice. It also means that when the numbers do improve, you can point at the specific ones you predicted.