The Blood Tests Worth Running If You Train Hard

If you train seriously and want blood work that tells you something, run a core panel four times a year and control the conditions tightly. Start with a complete blood count with reticulocytes, then ferritin with transferrin saturation and soluble transferrin receptor. Add hs-CRP and 25-hydroxyvitamin D. Add a comprehensive metabolic panel with creatine kinase and blood urea nitrogen. Add TSH with free T4, then total and free testosterone with SHBG. Finish with IGF-1, ApoB and HbA1c. Add a morning cortisol only if you are sampling at a fixed time. The panel matters less than the protocol and the time series. A single draw tells you almost nothing about a healthy trained person, because the width of population reference intervals swamps the within-person changes you care about.

Why the standard athlete panel underperforms

Commercial “performance panels” hand you thirty analytes and a green-yellow-red flag against a population interval derived from a mixed reference cohort. Most trained people sit inside those intervals on almost everything. The signal in athlete blood work is the deviation from your own history. Your position in a population distribution is not the signal.

Two things create that problem. First, the ratio of within-subject biological variation to between-subject variation is small for many analytes, which means your personal range occupies a narrow slice of the reference interval. Second, exercise itself moves many of these markers acutely, so a draw taken 18 hours after a hard session reflects the session and not your status. Reviews of biomarker use in sport make the same point repeatedly. Interpretation requires context on training load, timing and hydration, or the numbers mislead.¹²

The fix is boring and effective. Standardize collection, sample repeatedly, and compute your own reference interval.

The pre-analytical protocol

Plasma volume is the largest uncontrolled variable in athlete hematology. A day of heavy sweating contracts plasma volume and inflates hemoglobin and hematocrit. Several days of aerobic training expands it and deflates the same numbers. Work on the Athlete Biological Passport has shown that blood markers themselves can be used to estimate vascular volume and correct for this variance, which is a direct admission of how large the effect is.³

So fix the following and record them with every draw:

• Time of day: between 07:00 and 09:00, always the same hour within 30 minutes.
• Fasting: 10 to 12 hours, water allowed.
• Posture: seated and still for 15 minutes before venipuncture. Standing versus supine shifts plasma volume by several percent.
• Exercise: no session in the preceding 24 hours for the recovery panel, 48 hours if you did eccentric or novel work.
• Hydration: record morning body mass and urine specific gravity if you have a refractometer.
• Altitude and travel: note any exposure above 1500 m in the last 30 days.
• Menstrual cycle phase, if applicable, recorded as day since onset of menses.

Log these as columns next to the results. Without them the time series is not interpretable.

What each block of the panel is for

The hematology block is about oxygen transport and iron. One group is hemoglobin, hematocrit, reticulocyte percentage and mean corpuscular volume. The other is ferritin, transferrin saturation and soluble transferrin receptor. Read together, they separate true iron deficiency from the dilutional pseudo-anemia that is normal in endurance athletes. Ferritin is an acute-phase protein and rises with inflammation, so always interpret it alongside hs-CRP. If hs-CRP is above roughly 5 mg/L, the ferritin value is not usable as an iron store estimate that week.

The muscle damage and recovery block is creatine kinase and lactate dehydrogenase, plus aspartate aminotransferase and blood urea nitrogen. Creatine kinase peaks roughly 24 to 48 hours after eccentric loading and has enormous between-person variation, with some athletes running several thousand U/L without symptoms. Its value is entirely in the within-person trajectory. Work in soccer players tracking recovery across the microcycle found that these markers, sampled at fixed points relative to matches, describe recovery efficiency in a way single measurements cannot.⁴ Urea reflects protein catabolism and cumulative load rather than acute damage, so it moves on a slower timescale than creatine kinase.

The endocrine block covers testosterone with SHBG and IGF-1. It also covers TSH and free T4, plus optionally morning cortisol. The testosterone-to-cortisol ratio is popular and weakly informative, largely because cortisol has a steep diurnal slope and a within-day coefficient of variation that dwarfs any training effect you are trying to detect. If you sample cortisol at all, sample it at a fixed time and treat single values with skepticism. Sustained low free T4 with a low or low-normal TSH, or a testosterone trending down over multiple quarters alongside rising fatigue, is a pattern to bring to a physician rather than one to self-interpret.

The metabolic block is fasting glucose, HbA1c and ApoB. Add triglycerides and, if available, continuous glucose data. High training volume with high carbohydrate intake produces glucose patterns that look unusual against sedentary norms. Continuous glucose monitoring resolves this far better than a fasting spot value, because it shows post-session glycemia and overnight stability. Iron deficiency and high reticulocyte turnover both lower HbA1c independently of glycemia, which is another reason to read the panel as a whole.

Vitamin D, magnesium, and a full electrolyte set complete the picture. Reviews of the field position these routine analytes as the interpretable foundation, with metabolomic and proteomic panels layered on top once the basics are under control.⁵²

Computing your own reference interval

Here is the part most consumer panels skip. Once you have five or more draws under a standardized protocol, estimate your personal mean and within-subject coefficient of variation for each analyte. Then use a reference change value to decide whether a new result differs from your baseline:

RCV = 2^0.5 * Z * sqrt(CVa^2 + CVi^2)

CVa is analytical imprecision, which your lab publishes or you can request. CVi is your within-subject variation from your own series. Z is 1.96 for a two-sided 95% decision. For ferritin, with analytical CV near 5 percent and within-subject CV commonly around 15 percent, the RCV lands near 44 percent. A ferritin move from 40 to 50 ng/mL is noise. A move from 40 to 20 ng/mL is signal.

Practically, keep results in one tidy CSV. Use columns for date, analyte and LOINC code. Add value, unit and lab, plus the covariates above. Long format, one row per measurement. Then fit a simple model per analyte in R or Python: a loess or state-space smoother over time, with the plasma volume proxies as covariates for hemoglobin and hematocrit. Anti-doping science took exactly this route with Bayesian adaptive models that flag individual deviations against a personalized prior rather than a population cutoff.⁶ The statistical machinery is public, and the same logic applies to training monitoring.

Two failure modes to avoid. Do not mix labs and assay platforms without noting it. Ferritin, testosterone and IGF-1 assays disagree by clinically meaningful margins across vendors. And do not sample opportunistically after a hard block, because you will build a baseline out of stress responses and then misread your rested state as abnormal.

When this needs a clinician

Blood work is measurement. Any persistently abnormal result belongs with a physician. That includes anemia, ferritin below the range your lab flags and thyroid abnormalities. It also includes a testosterone that keeps declining or elevated liver enzymes that do not resolve with rest. Iron supplementation in particular should never be self-directed on a ferritin value alone, because iron overload is a real risk and hereditary hemochromatosis is common enough to matter. Take the data and the covariate log to the appointment.

Questions people also ask

How often should I test? Quarterly for the full panel, with an extra draw at the end of a heavy block if you are tracking recovery specifically. Monthly ferritin during a deliberate iron repletion period, if a clinician is managing it.

Is a single panel useless? It is limited, though not useless. It can catch a frank abnormality. It cannot tell you whether you are adapting or accumulating fatigue, because that requires a personal baseline.

Why does my hemoglobin drop during heavy training? Usually plasma volume expansion, which dilutes red cell concentration without reducing total red cell mass. Reticulocytes and iron markers help distinguish dilution from real loss.³

Should I track creatine kinase routinely? Only against your own history and at a fixed interval after a known session type. Absolute values compared to population ranges are close to meaningless for athletes.⁴

Footnotes

1. Elaine C. Lee, Maren S. Fragala, Stavros A. Kavouras, et al. Biomarkers in Sports and Exercise: Tracking Health, Performance, and Recovery in Athletes. Journal of Strength and Conditioning Research, 2017. https://doi.org/10.1519/jsc.0000000000002122 ↩

2. Mario Muñoz-López, Gonzalo Quesada-Fernández, Edgar Simón Sancho-Haro, et al. From Routine Blood Tests to Metabolomics: A Contextual Framework for Interpreting Biomarkers of Training Load, Recovery, and Metabolic Stress in Athletes. Metabolites, 2026. https://doi.org/10.3390/metabo16070483 ↩ ↩²

3. Louisa M. Lobigs, Pierre‐Edouard Sottas, Pitre C. Bourdon, et al. A step towards removing plasma volume variance from the Athlete’s Biological Passport: The use of biomarkers to describe vascular volumes from a simple blood test. Drug Testing and Analysis, 2017. https://doi.org/10.1002/dta.2219 ↩ ↩²

4. Anna Nowakowska, Dorota Kostrzewa-Nowak, Rafał Buryta, et al. Blood Biomarkers of Recovery Efficiency in Soccer Players. International Journal of Environmental Research and Public Health, 2019. https://doi.org/10.3390/ijerph16183279 ↩ ↩²

5. Iñigo San-Millán. Blood Biomarkers in Sports Medicine and Performance and the Future of Metabolomics. Methods in Molecular Biology, 2019. https://doi.org/10.1007/978-1-4939-9236-2_26 ↩

6. Raphael Faiss, Jonas Saugy, Martial Saugy. Fighting Doping in Elite Sports: Blood for All Tests!. Frontiers in Sports and Active Living, 2019. https://doi.org/10.3389/fspor.2019.00030 ↩