Can THC or CBD Thicken Whole Blood? What a 24-Donor Laboratory Experiment Measured

Short answer: In a laboratory experiment involving venous whole blood donated by 24 healthy adults, adding specified concentrations of delta-9-tetrahydrocannabinol (THC) or cannabidiol (CBD) to blood samples increased measured viscosity at 25–100 ng/mL. Two THC:CBD-containing extracts produced still higher readings and increased the laboratory measure of elasticity. Microscopy also showed altered red-cell shapes after higher-concentration treatments. These were changes in blood treated outside the body—not blood changes demonstrated in people after using cannabis. The experiment measured mechanical behavior after deliberate laboratory spiking; it did not measure a user’s circulating THC concentration, clinical blood flow, clotting, vaso-occlusive events, driving performance, or individual impairment.

The primary source is James and colleagues, “The in vitro effect of delta-9-tetrahydrocannabinol and cannabidiol on whole blood viscosity, elasticity and membrane integrity,” Journal of Cannabis Research (2022), doi:10.1186/s42238-022-00126-z. The distinctions below follow the complete original methods, results, micrographs, and discussion rather than treating the abstract’s clinical interpretation as experimental proof.

What question did the experiment actually answer?

It asked whether laboratory-applied THC, CBD, or combined extracts would change the flow resistance, recoverable deformation, and observable red-blood-cell morphology of donated human whole blood. That is a legitimate ex vivo rheology question. It is not a human cannabis-exposure trial. Participants supplied blood; the investigators added cannabinoids to aliquots after collection. No participant was administered THC or CBD as part of the experiment.

Blood viscosity describes resistance to deformation and flow under specified test conditions. In an oscillatory viscoelastic system, an elasticity-related component describes stored mechanical response. Neither is a stand-alone measurement of clot formation, oxygen delivery, stroke risk, or acute intoxication. This matters because an unusually high device reading can be an experimentally meaningful result without being a validated predictor of what happens inside a living patient’s circulation.

Who donated the blood, and what was the true sample size?

The investigators recruited 24 healthy adult nonsmokers: 14 women and 10 men, ages 21–42, at the University of the West Indies, Mona, Jamaica. All were reported to have hemoglobin AA by electrophoresis. The study excluded reported cannabis use, circulatory disease, certain medication or illicit-drug exposure, and recent transfusion. The source did not impose a clearly defined lookback interval for every exclusion and did not require detailed reporting of all other alcohol, tobacco, or herbal use.

Each volunteer contributed a venous blood specimen. Multiple treated portions and multiple instrument conditions do not create additional independent donors. The figures describe n=24 for plotted responses, although some reported pairwise statistical tests have smaller degrees of freedom; the source does not fully reconcile every analysis-specific denominator. It would be wrong to describe 24 people as 24 cannabis users, 24 clinical adverse events, or hundreds of independently dosed patients.

How were the samples collected and treated?

Investigators drew approximately 8 mL of antecubital venous blood per person, placing it into potassium-EDTA anticoagulant tubes. The blood was held near 25°C before testing. Cannabinoid stock preparations used ethanol; investigators prepared nominal THC-alone and CBD-alone mixtures at 2.5, 25, 50, and 100 ng/mL and compared them with an untreated control. Treated portions were mixed, held for approximately 20 minutes near room temperature, and subsequently analyzed.

The mixtures are best understood as nominal concentrations introduced by the laboratory. The paper does not report a separate quantitative chromatographic confirmation of parent THC or CBD in each final whole-blood aliquot. Thus “100 ng/mL THC” identifies the investigators’ target addition, not a verified whole-blood concentration observed after a person inhaled or swallowed THC. Likewise, the study’s concentration series was not a pharmacokinetic time course.

What instrument measured viscosity and elasticity?

The original authors used a Vilastic Scientific BioProfiler to obtain viscoelastic measurements from whole blood at native hematocrit. The stated principal measurement conditions were approximately 37°C, oscillation frequency 2 Hz, shear rate 62.8 s⁻¹, and strain 5. Hematocrit was determined using a capillary centrifugation method. The paper’s plotted results are expressed in millipascal-seconds (mPa·s), with errors described as standard errors of the mean—not standard deviations or confidence intervals.

Rheology is dependent on test conditions. Measurements taken at this shear rate and oscillation frequency should not be casually equated with a clinical reference interval obtained using another rheometer or a different shear regime. Similarly, a test conducted under controlled temperature after short in-vitro incubation cannot demonstrate what blood viscosity would be during changing metabolism, binding, circulation, and tissue distribution in a living person.

How much did THC alone change measured whole-blood viscosity?

The THC-treated series began with a mean control viscosity of 3.74 ± 0.21 mPa·s. At a nominal 25 ng/mL THC, the reported mean rose to 6.45 ± 0.36 mPa·s; at 50 ng/mL it was 8.07 ± 0.43 mPa·s; and at 100 ng/mL it reached 11.60 ± 1.12 mPa·s. The authors reported a significant concentration effect for the THC series, F(1.36, 20.40)=53.71, p<.001.

The lowest 2.5 ng/mL condition did not significantly change viscosity relative to control. That negative finding prevents an unsupported claim that every tested amount increased viscosity. The means also do not establish a clinical cutoff at 25 ng/mL: the experiment selected concentrations and identified statistical group-level differences in a particular apparatus, not a dangerous individual blood level in cannabis consumers.

How did CBD alone compare with THC?

CBD-treated aliquots showed a related trend. The CBD control was 3.77 ± 0.17 mPa·s; at 25 ng/mL it was 5.46 ± 0.24 mPa·s; at 50 ng/mL 7.51 ± 0.42 mPa·s; and at 100 ng/mL 9.91 ± 1.10 mPa·s. The reported CBD repeated-measures effect was F(1.27, 25.47)=27.39, p<.001, with no significant change at the lowest 2.5 ng/mL concentration.

THC group means exceeded the corresponding CBD means across the higher tested levels, but the paper did not establish a universally reliable superiority of THC over CBD at every concentration. Some between-compound comparisons are reported with statistics that deserve independent checking, particularly one printed t statistic and p value that do not agree arithmetically. The safest interpretation is the demonstrated within-series concentration trend, while preserving uncertainty about every specific between-drug contrast.

Evidence table 1: the separate-cannabinoid viscosity measurements

Values are original paper’s descriptive measurements. “Not numerically tabulated” avoids reverse-engineering a precise 2.5-ng/mL mean from a plot. This table cannot be used to predict a person’s viscosity from an isolated THC concentration or to convert a plasma result into a legal whole-blood result.

What happened when THC and CBD were combined in extracts?

The authors also compared two mixed preparations with nominal 50 ng/mL THC. One had a stated 4:1 THC:CBD ratio, corresponding to 12.5 ng/mL CBD; the other had a 1:1 ratio, corresponding to 50 ng/mL CBD. The resulting mean viscosity readings were 21.33 ± 2.17 mPa·s and 21.76 ± 1.88 mPa·s, compared with an extract-series control of 3.82 ± 0.81 mPa·s. The reported overall comparison was F(2,38)=53.13, p<.001.

Those mixed-preparation readings were much higher than the separate-THC or separate-CBD results at 50 ng/mL. Yet the two extracts were not significantly different from each other. Greater viscosity under these experimental conditions is a defensible observation. Saying that a precisely quantified biochemical “entourage synergy” was established is stronger than the design permits: preparation constituents, solvent matching, and combined-concentration effects were not isolated with a complete factorial vehicle-controlled dose-response experiment.

Did the blood become more elastic as well?

The investigators reported no significant increase in their elasticity measurement with THC alone or CBD alone across the examined concentrations. The mixed extracts behaved differently. Their reported elasticity means were 4.41 ± 0.57 mPa·s for 4:1 and 5.15 ± 0.76 mPa·s for 1:1, versus 1.87 ± 0.51 mPa·s for control, with F(1.53,30.50)=7.88 and p=.004.

The two mixed ratios were again not significantly different from one another. A critical documentary detail is that the original article’s figure captions for its separate-concentration elasticity plot and mixed-extract elasticity plot appear interchanged: the axis labels and data layout clarify which graph actually addresses each question. Reporting the numeric result with the specified condition is more reliable than quoting a figure number without checking its plotted axes.

Evidence table 2: extracts, elasticity, and true negative findings

Means are accompanied by the original SEM. Reported global tests do not justify inferring a cutoff, a precise clinical increase, or an individual likelihood of thrombosis. The instrument did not test live vascular function.

What did the red-blood-cell microscopy show?

Investigators stained blood films with Wright’s stain and examined erythrocytes microscopically at ×100 oil-immersion magnification. Untreated control films showed typical biconcave erythrocyte appearance. At the lower 2.5 and 25 ng/mL separate THC or CBD preparations, investigators described morphology resembling the control. At nominal 50 ng/mL, the displayed THC and CBD samples showed burr-cell-like projections and blister or membrane abnormalities. At 100 ng/mL, images showed more fragmentation and degeneration.

Combined extracts were described as producing conspicuous red-cell agglutination or severe damage, with the 4:1 extract image displaying very few intact erythrocytes. The micrographs are striking, but the paper did not provide a validated numerical percentage of damaged cells, blinded scoring, quantitative hemolysis biomarker, or clinical microvascular outcome. A representative photograph is not the same as a measured population incidence of red-cell rupture.

There is also an important distinction between cellular morphology after drug addition in a tube and hemolysis in a living human. The former was examined; the latter was not measured. The appearance of laboratory-agglutinated erythrocytes does not itself prove that someone using cannabis will experience an acute stroke, sickle-cell crisis, thrombosis, or other event.

Why the specimen is correctly called venous whole blood—but the drug level is not a measured BAC-style result

The authors truly drew venous whole blood. That identifies the physical matrix subjected to mechanical testing. The cannabinoids, however, were added to that material at laboratory-selected nominal concentrations. The study did not analyze a sample from a person after cannabinoid consumption and did not confirm each aliquot’s resulting parent-drug concentration using an independent forensic quantitative assay. These two statements must appear together to avoid a misleading phrase such as “patients with a blood THC of 50 ng/mL developed thick blood.”

For other forensic tests, a properly collected venous whole-blood THC result and a separated plasma or serum THC result refer to different matrices and should not be interchanged without supporting research. Urine THC-COOH generally addresses a metabolite in a different matrix, not active THC in whole blood. Oral fluid, exhaled breath, hair, and vitreous are different specimens again. See the related explanation of biological matrices and limits on interpreting concentrations.

Can the concentration series be equated with cannabis doses or actual impairment?

No. The 2.5–100 ng/mL figures represent what the investigators placed into collected blood, not a person’s absorbed inhaled or oral dose. In ordinary pharmacokinetics, distribution among plasma, erythrocytes, proteins, and tissues changes with time. Drug administration, body composition, collection time, and analytical preparation all matter. A target concentration mixed into a tube is not synonymous with the concentration that follows smoking a product of specified THC percentage or taking a prescribed oral dose.

The paper mentions earlier literature concerning plasma THC and subjective feeling states as rationale for selecting concentrations. That does not convert this experiment into a human exposure-effect trial. No participants rated intoxication after research dosing, completed a driving simulator, drove a car, performed field sobriety tests, or supplied contemporaneous clinical symptom data in response to cannabinoids. No blood pressure, clotting event, or active organ disease endpoint was observed. An individual impaired-driving inference or event attribution cannot be derived from this rheometer’s group means.

What are the chief design and analytical limitations?

First, exposure occurred ex vivo. Whole blood was collected once and manipulated outside the donor. The human body’s clearance, tissue distribution, vascular endothelium, oxygenation, immune responses, and compensatory flow were not represented. The blood’s physical behavior after static incubation is not proof of a sustained living effect.

Second, controls need closer scrutiny. The authors describe ethanol-based cannabinoid stocks and state that untreated blood acted as control. They do not describe a fully matched final-solvent concentration and extract-excipient panel establishing that every altered measurement was attributable only to THC or CBD. The mixed preparations are identified mainly by cannabinoid ratios. That leaves important questions about vehicle, other components, and interaction mechanisms.

Third, measurements are method-bound. Venous EDTA blood was incubated around 25°C and tested under a specific oscillatory/shear setup at 37°C. Hematocrit and sample handling can change rheology. Their group means and SEM are not validated diagnostic boundaries or uncertainty distributions for unrelated clinical devices.

Fourth, clinical denominators and statistics matter. The group contained 24 HbAA donors, not persons with sickle cell disease. Some pairwise t tests use fewer degrees of freedom, and at least one stated t/p pairing warrants verification. The figure-caption labels for the elasticity panels are also inconsistent with the graphics. A conscientious evaluation should prefer the actual numeric result and its endpoint over unwarranted precision.

Fifth, morphology was descriptive. Photographs document abnormalities in selected blood films, but no numerical hemolysis assay, blinded cell-count distribution, vascular occlusion endpoint, or prospective clinical observation was supplied. Statements about clinical crises are biological hypotheses, not outcomes measured in this original experiment.

What could a forensic or civil-litigation reviewer legitimately say?

The reviewable proposition is narrow: under the described laboratory conditions, added cannabinoids changed mechanical properties of donated human venous whole-blood aliquots, with clear higher-concentration viscosity effects, mixed-extract elasticity effects, and visible cellular abnormalities in selected stained films. That proposition can be discussed as mechanistic or hypothesis-generating evidence about blood rheology.

It cannot establish that a particular driver, decedent, worker, or patient had altered blood flow or became impaired because of THC or CBD. Nor does it show that cannabis caused a sickle cell crisis, stroke, coagulation abnormality, or traffic collision. Individual opinions would need appropriate clinical records, actual toxicological specimen results, history, timing, validated analyses, competing causes, and a scientifically relevant causal model.

For a case relying on a real drug finding, distinguish a targeted instrumental result from a nonspecific screen and obtain method-validation and quality-control materials; see confirmatory drug testing principles. Collection, anticoagulant, time, and storage also matter; see forensic specimen collection considerations.

What documents would resolve the most important unanswered laboratory questions?

Before reproducing the viscosity conclusions in litigation, a reviewer would reasonably seek individual donor-level rheometer outputs, hematocrits, sample randomization and replicate handling, the final solvent content of each treated aliquot, a genuine vehicle-matched control, and a complete constituent analysis of each extract. The original raw data would allow checking the reported repeated-measures and pairwise statistics as well as the number of donors contributing to each comparison.

Additional evidence would be required to characterize hemolysis quantitatively: matched microscopy across donors, standardized morphology counts, free hemoglobin or another appropriate biological damage marker, and investigation of whether observed effects persist under physiologically relevant exposure conditions. Those proposals identify what was not established; they are not results claimed by James and colleagues.

Frequently asked questions

1. Did 24 people take THC or CBD and develop thicker blood?

No. Twenty-four healthy nonsmokers donated venous blood. Researchers subsequently spiked portions of the collected specimens in a laboratory. There was no administered cannabis trial in these donors and no measurement of their response after taking a product.

2. Was THC measured in whole blood after a person used cannabis?

No. Whole blood was genuinely the tested physical specimen, but the drug numbers were target laboratory additions. The paper did not report forensic confirmation of post-use THC, CBD, 11-hydroxy-THC, or THC-COOH in participant whole blood.

3. Did every THC or CBD concentration raise viscosity significantly?

No. The study reported no significant change at its lowest 2.5 ng/mL addition. Higher concentrations were associated with higher mean viscosity under the specified laboratory conditions. Those groups do not define a safe or unsafe person-level concentration.

4. Did combining THC and CBD establish a clinically harmful synergistic effect?

The two mixed extracts yielded much higher mean laboratory viscosity and elasticity readings than control and greater viscosity than THC or CBD alone at the chosen comparison. But the paper did not fully isolate all preparation constituents, final solvent conditions, or in-vivo mechanisms. Clinical synergy and patient harm therefore remain unproven.

5. Did microscopy prove clinically significant hemolysis or blood clotting?

No. The original stained films showed changes in shape, fragmentation, and agglutination after selected treatments. They did not quantify hemolysis with a validated clinical assay or measure thrombosis, vascular occlusion, or stroke.

6. Can these values be compared directly with plasma THC, urine THC-COOH, or breath readings?

No. The results are viscoelastic parameters measured in treated whole blood, expressed in mPa·s. They are not cannabinoid laboratory concentrations in a living patient. Drug concentrations reported for different biological matrices require their own methods and cannot be transposed by analogy.

7. Does this study prove cannabis caused an individual’s impaired driving or cardiovascular event?

No. Neither driving performance nor an individual cardiovascular event was tested. The experiment offers laboratory mechanistic evidence only, not a validated threshold or a case-specific causal conclusion. A careful interpretation must remain confined to the conditions and endpoints actually measured.

Original primary research

James TR, Richards AA, Lowe DA, Reid WA, Watson CT, Pepple DJ. The in vitro effect of delta-9-tetrahydrocannabinol and cannabidiol on whole blood viscosity, elasticity and membrane integrity. Journal of Cannabis Research. 2022;4:15. https://doi.org/10.1186/s42238-022-00126-z. Conclusions above are constrained by the complete original study’s design, numerical figures, and directly observed endpoints.