What the Most Comprehensive Blood Test Measures

Oak
blood-tests biomarkers proteomics lab-data measurement
A laboratory instrument holding a vertical glass plate of thousands of individually glowing tiny wells, fed by a robotic arm of fine glass tips.

There is no single blood test that checks everything, and the panels marketed that way are usually the same forty or so clinical chemistry analytes bundled under a new name. The most comprehensive blood measurement available to an individual today is a stack: a complete blood count with differential, a comprehensive metabolic panel, a lipid panel with apolipoprotein B and lipoprotein(a), thyroid and sex hormones, iron studies, inflammatory markers, vitamins and minerals, and then a research-grade affinity proteomics run measuring thousands of plasma proteins at once. The clinical portion costs roughly $200 to $600 out of pocket; the proteomics portion is where the information density jumps by three orders of magnitude, and it is not part of any standard panel. This page describes what each layer measures, what it cannot tell you, and how to get the output as data rather than a PDF.

What a standard “full panel” contains

Every consumer panel sold as comprehensive is assembled from the same building blocks, and it helps to know their contents precisely so you can tell whether a $99 panel and a $499 panel differ in substance.

A complete blood count (CBC) with differential reports around twenty values: red cell count, hemoglobin, hematocrit, MCV, MCH, MCHC, RDW, platelet count, MPV, and the white cell count broken into neutrophils, lymphocytes, monocytes, eosinophils, and basophils as both percentages and absolute counts. These are cheap, reproducible, and more informative in combination than in isolation. Ratios derived from the differential, such as neutrophil-to-lymphocyte ratio, carry signal about systemic inflammation and have been studied as discriminators in oncology contexts, though with modest effect sizes and substantial overlap between groups.1

A comprehensive metabolic panel (CMP) is fourteen analytes: glucose, BUN, creatinine, eGFR, sodium, potassium, chloride, CO2, calcium, total protein, albumin, total bilirubin, alkaline phosphatase, AST, and ALT. It covers kidney function, liver enzymes, electrolytes, and one fasting glucose reading. A lipid panel adds total cholesterol, HDL-C, triglycerides, and calculated or direct LDL-C.

Beyond these, the additions that carry the most information per dollar are apolipoprotein B (a direct count of atherogenic particles rather than the cholesterol carried inside them), lipoprotein(a) (largely genetically determined, so once in a lifetime is enough), hemoglobin A1c and fasting insulin, high-sensitivity C-reactive protein, ferritin with transferrin saturation, TSH with free T4 and free T3, and 25-hydroxyvitamin D. Vitamin D deserves a note on interpretation: 25(OH)D is the accepted status marker because of its long half-life and its reflection of both dietary intake and cutaneous synthesis, whereas 1,25-dihydroxyvitamin D is tightly regulated and can appear normal or high in deficiency.2 People order the wrong one constantly.

That whole stack is perhaps 60 to 80 numbers. It is the floor, not the ceiling.

Where the information increases: plasma proteomics

The plasma proteome is the largest accessible molecular readout in blood, because every tissue in the body leaks or secretes protein into circulation, and the resulting mixture is a diluted summary of systemic physiology. This argument was laid out clearly two decades ago, along with the central technical obstacle: dynamic range.3 Albumin and immunoglobulins constitute most of the protein mass, and the interesting low-abundance analytes sit ten orders of magnitude below them.

Two platform families solve this differently. Affinity proteomics (Olink’s proximity extension assay, SomaLogic’s aptamer-based SomaScan) reads 3,000 to 11,000 protein targets from roughly 150 microliters of plasma, reported as relative abundance: Olink gives Normalized Protein eXpression values on a log2 scale, SomaLogic gives relative fluorescence units. Mass spectrometry, by contrast, gives absolute identification and is not limited to a predefined panel, but typically quantifies only a few hundred to a few thousand proteins in undepleted plasma. We recommend affinity proteomics as the default for an individual baseline. You get far more targets per sample, the coefficients of variation are usually in the 5 to 15 percent range for well-behaved assays, and the output is a clean matrix rather than a spectral processing project.

The tradeoff is real. Affinity measurements are relative, not absolute, so a single timepoint tells you where you sit in a reference distribution and little else. Aptamer and antibody binding can be disrupted by a coding variant in the epitope region, which produces an artifactual shift in one protein that looks biological and is not. This is why pairing proteomics with whole-genome sequencing matters: you can check whether a low reading sits on top of a missense variant in that gene. And batch effects between runs are large enough that comparing your 2026 sample against your 2028 sample requires bridging samples or careful normalization, not naive subtraction.

The analytes that do not belong in a general panel

A comprehensive panel should measure physiology, not screen for rare disease in an asymptomatic person. Several categories that sound appealing fall apart on base rates.

Circulating tumor DNA and multi-analyte cancer detection are the clearest case. The CancerSEEK study combined mutations at 61 amplicons of cell-free DNA with eight protein markers and reported sensitivity of roughly 70 percent across eight cancer types at greater than 99 percent specificity, with sensitivity varying from about 33 percent in breast cancer to nearly 98 percent in ovarian cancer.4 Those are strong numbers for a research assay, and still, at population prevalence, most positives in an asymptomatic person are false. Related work extends the idea with cell-free RNA, methylation, and fragmentomics layered together, which improves discrimination in cohorts where cases are enriched.5 Circulating microRNAs show similar behavior: measurable differences between patients and controls, insufficient positive predictive value for undirected screening.6 Extracellular vesicle biomarkers face an additional stoichiometric limit, because the number of target molecules per vesicle and the fraction of vesicles originating from any given tissue set a hard floor on detectability in a few milliliters of plasma.7

The same logic applies to immune monitoring markers. Peripheral blood immune biomarkers are being developed to track response to checkpoint blockade, and they are promising in that setting because there is a treated patient and a specific question.8 Outside it, they are noise. Multiplex panels for neurodegeneration follow the same pattern: recent work identified blood and urine markers that separate people who later received a Parkinson’s diagnosis from controls years before clinical presentation, which is scientifically important and still a cohort-level finding rather than a personal test.9 Any result that suggests a disease process requires a physician to interpret and confirm; none of these assays are self-interpretable, and acting on one without clinical evaluation is how people end up with unnecessary imaging and biopsies.

Getting the data, not the PDF

The practical failure mode of comprehensive testing is that you receive a styled report with green and red arrows and no underlying file. Insist on the following.

For clinical chemistry, request the HL7 or FHIR Observation resources from the lab portal, or at minimum a CSV with LOINC codes, numeric results, units, and the reference interval used. LOINC codes matter because “vitamin D” from two labs may be different analytes, and because joining panels over five years by test name string will fail. Record the collection timestamp, fasting status, and time since last exercise. Strenuous exercise within 48 hours raises creatine kinase, ALT, and AST enough to be misread as liver signal.

For proteomics, the deliverable you want from Olink is the NPX file with sample and assay QC flags retained, plus the limit-of-detection column per assay; from SomaLogic, the ADAT file, which contains RFU values along with the normalization scale factors applied. Do not accept a version with failed assays silently dropped. In R, OlinkAnalyze::read_NPX() and SomaDataIO::read_adat() handle these directly. Filter to assays where your value exceeds the LOD, then work on the log scale.

For longitudinal work, the design decision that matters most is standardizing the pre-analytical conditions: same tube type, same fasting duration, same time of day, same lab. Cortisol, iron, and testosterone have diurnal swings larger than most effects you would hope to detect. Two samples drawn at 7am and 4pm are not comparable, and no statistical correction will rescue them.

Questions people also ask

What is a full blood test for everything? It does not exist as a single test. The broadest routinely available combination is CBC with differential, CMP, lipids with ApoB, thyroid and sex hormones, iron studies, hsCRP, HbA1c, insulin, and vitamins D and B12, which totals roughly 60 to 80 analytes. Adding affinity proteomics raises that to several thousand protein measurements.

How much does a full blood panel cost? A comprehensive clinical panel from a direct-to-consumer lab runs about $200 to $600 without insurance. Research-grade plasma proteomics is typically $1,000 to $3,000 per sample depending on the platform and target count, and is usually purchased as part of a larger program rather than à la carte.

What is the most detailed blood test? By raw analyte count, an 11,000-target aptamer panel or a 5,400-target proximity extension assay. By clinical decision-making value for a healthy person, the boring chemistry panel plus ApoB, Lp(a), and HbA1c still carries more actionable weight, because those markers have decades of outcome data behind their reference ranges.

Can doctors tell everything from a full blood test? No. Blood reflects what tissues release into circulation, which means organ-specific processes are visible only when they produce a detectable protein or nucleic acid signal above background.3 Structural findings, early localized disease, and many neurological and psychiatric conditions leave no reliable blood signature at all.

Oak builds longitudinal molecular profiles of individuals: whole-genome sequencing, RNA sequencing, proteomics, blood biomarkers, and continuous glucose data, integrated into one model of you. Build your profile.

Footnotes

  1. Neşe Bülbül, Deniz Gülnihal, Rukiye Çiftçi, et al. Preoperative Thyroid Hormone Ratios and Complete Blood Count (CBC)-Derived Biomarkers for Differentiating Benign Thyroid Nodules from Papillary Thyroid Carcinoma: A Retrospective Single-Center Case–Control Study. Journal of Clinical Medicine, 2026. https://doi.org/10.3390/jcm15155913 ↩

  2. Joseph E Zerwekh. Blood biomarkers of vitamin D status. The American Journal of Clinical Nutrition, 2008. https://doi.org/10.1093/ajcn/87.4.1087s ↩

  3. Lance A. Liotta, Mauro Ferrari, Emanuel Petricoin. Clinical proteomics: Written in blood. Nature, 2003. https://doi.org/10.1038/425905a ↩ ↩2

  4. Joshua D. Cohen, Lu Li, Yuxuan Wang, et al. Detection and localization of surgically resectable cancers with a multi-analyte blood test. Science, 2018. https://doi.org/10.1126/science.aar3247 ↩

  5. Yuhuan Tao, Shaozhen Xing, Shuai Zuo, et al. Cell-free multi-omics analysis reveals potential biomarkers in gastrointestinal cancer patients’ blood. Cell Reports Medicine, 2023. https://doi.org/10.1016/j.xcrm.2023.101281 ↩

  6. Helen M. Heneghan, Nicola Miller, Aoife J. Lowery, et al. Circulating microRNAs as Novel Minimally Invasive Biomarkers for Breast Cancer. Annals of Surgery, 2010. https://doi.org/10.1097/sla.0b013e3181cc939f ↩

  7. Nataša Zarovni, Danilo Mladenović, Dario Brambilla, et al. Stoichiometric constraints for detection of EV‐borne biomarkers in blood. Journal of Extracellular Vesicles, 2025. https://doi.org/10.1002/jev2.70034 ↩

  8. Ioannis P. Trontzas, Konstantinos N. Syrigos. Immune Biomarkers for Checkpoint Blockade in Solid Tumors: Transitioning from Tissue to Peripheral Blood Monitoring and Future Integrated Strategies. Cancers, 2025. https://doi.org/10.3390/cancers17162639 ↩

  9. Shuo Gao, Zheng Wang, Yuanfeng Huang, et al. Early detection of Parkinson’s disease through multiplex blood and urine biomarkers prior to clinical diagnosis. npj Parkinson’s Disease, 2025. https://doi.org/10.1038/s41531-025-00888-2 ↩