Does Severe Kidney Disease Increase Plasma THC After the Same Cannabis Dose? A 49-Person Study

Answer: In a controlled-dose, nonrandomized study of 49 adults, people with stage 4 or nondialysis stage 5 chronic kidney disease (CKD) had higher measured plasma THC and cannabidiol (CBD) exposure over 24 hours than healthier controls after the same two sprays of Sativex. Median THC area under the concentration–time curve (AUC) was 2.76 h·ng/mL in 20 controls, 4.16 in 20 stage 4 patients, and 4.31 in nine stage 5 patients. Both CKD groups differed from controls, but not significantly from each other. The study did not measure forensic whole-blood THC, validate an impairment cutoff, or show that kidney function independently caused the higher THC AUC after age, sex, and body mass index were accounted for.

The original human experiment was reported by Sønderskov and colleagues in Kidney International Reports (2025): Pharmacokinetics and Side Effects of Δ9-Tetrahydrocannabinol and Cannabidiol in Patients with Different Stages of CKD. Its forensic value is not a universal kidney-disease correction factor. It is evidence that the same nominal cannabinoid dose can produce different plasma concentration histories in clinically different populations—and that the endpoint, specimen and statistical adjustment matter.

What question did the investigators actually test?

They asked whether adults with severely reduced kidney function experience different systemic cannabinoid pharmacokinetics and short-term symptoms after a standard, fixed oromucosal THC/CBD exposure. The primary endpoint was THC AUC from administration through 24 hours. Secondary endpoints included AUC, peak concentration, peak timing, urinary metabolite excretion and self-reported symptoms. This was a single-dose clinical pharmacokinetic comparison, not a roadside cannabis study, randomized kidney-function intervention or clinical trial of chronic cannabis efficacy.

AUC summarizes cumulative plasma concentration × time over the sampled interval. A higher AUC does not mean that every plasma sample was higher, that the administered dose differed, or that the person was impaired for that entire period. AUC of 4 h·ng/mL is not a measured single-sample THC concentration of 4 ng/mL. Confusing these units can invalidate an expert comparison before any discussion of pharmacology begins.

Who was tested—and who was not?

Investigators recruited 66 people and analyzed 49 who supplied study data for more than six hours: 20 controls with estimated glomerular filtration rate (eGFR) above 60 mL/min/1.73 m², 20 with stage 4 CKD (eGFR greater than 15 and no greater than 30), and nine with nondialysis stage 5 CKD (eGFR no greater than 15). Dialysis patients were a separately envisaged category, but their results were not included in this publication. The nine stage 5 participants cannot be presented as a large representative cohort of all people with kidney failure.

The groups were demonstrably different beyond eGFR. Median ages were 27, 64 and 68 years for controls, stage 4 and stage 5 respectively. Median counts of regular medications were one, eight and seven. These distinctions matter because age, comorbidities, concurrent drugs, body composition, hepatic enzyme activity and treatment histories can alter cannabinoid exposure or perceived effects. The original analysis acknowledged that a more closely matched comparison would strengthen causal interpretation.

What dose, route and sampling schedule were used?

After seven hours of fasting, each participant received two oromucosal Sativex sprays totaling 5.4 mg THC and 5.0 mg CBD, directed inside the cheek. Swallowing was not restricted. Therefore the trial cannot separately quantify how much of each dose crossed oral mucosa versus entered the gastrointestinal tract. It also cannot be extrapolated without validation to smoking, vaping, an edible containing a very different dose, or repeated long-term cannabinoid use.

Investigators sampled blood and asked symptom questions at baseline and 14 postdose times through 24 hours, including 15, 30 and 45 minutes; one, two, four, six, eight, twelve and twenty-four hours. Four of 735 planned blood samples were missing. Urine was collected as a separate 24-hour cumulative specimen. These repeated samples were collected from the same 49 participants; they are not hundreds of independent people. Fasting and fluid restrictions also standardized circumstances that may differ substantially in real cases.

What did the laboratory directly measure?

Blood was collected in oxalate tubes, centrifuged, and the separated plasma was analyzed by ultra-high-performance liquid chromatography–tandem mass spectrometry. Analytes included parent THC, CBD, active 11-hydroxy-THC (THC-OH), inactive THC-COOH, THC-COOH-glucuronide, and the acidic plant constituents THCA-A and CBDA. In this method the reported plasma lower limits of quantification were 0.025 ng/mL for THC, 0.03 for CBD, 0.1 for THC-OH and THC-COOH, and 0.2 for THC-COOH-glucuronide. The proposed 7-hydroxy-CBD measurement was not performed because a reference standard was unavailable.

The paper also reports total excreted masses of selected cannabinoids in urine over 24 hours. That is a different physical matrix and a different question from plasma concentrations. It does not supply forensic venous whole-blood parent THC results or a validated conversion from plasma to whole blood. An investigator must not silently replace plasma with serum or whole blood, or replace a metabolite with parent THC, merely because the values share the unit ng/mL.

What were the parent THC and CBD results?

The original Table 2 gives median AUC, Cmax and Tmax for each group, using a Kruskal–Wallis overall test with corrected group comparisons. Both CKD groups had significantly higher 24-hour plasma AUC for parent THC than controls. The CBD contrast was larger in relative terms. However, the two CKD groups did not show a significant parent-THC AUC difference from each other; it would be incorrect to infer a demonstrated dose-response increase in THC exposure from stage 4 to stage 5.

All entries are medians, not individual predictions; AUC and Cmax refer to different quantities. The study’s full tables provide interquartile ranges and individual minima and maxima. For THC AUC the interquartile ranges were 1.77–3.48, 3.35–5.28 and 3.16–5.42 h·ng/mL, respectively—an immediate warning that distributions overlap. For CBD AUC they were 0.49–1.81, 2.49–3.66 and 2.22–4.66. The original printout writes some very small p values as “0.000”; proper reporting is p < .001, not zero probability.

How should the peak concentration and time-to-peak changes be interpreted?

A THC median peak of 1.09 ng/mL in stage 5 versus 0.62 ng/mL in controls is a plasma peak observed in this scheduled experiment. It is not a person-specific concentration at an unsampled roadside event. Repeated profiles show when the median participant reached a maximum; a single case specimen cannot reveal that its donor was at their own peak without additional timed data.

The stage 5 THC median Tmax was one hour, compared with 1.75 hours in controls and 1.50 hours in stage 4. Original corrected comparisons distinguished the stage 5 group from the others, but did not establish the exact cause of earlier peaks. The oromucosal-plus-swallowed dosing route, delayed sampling, heterogeneity and participant characteristics are plausible contributing dimensions, not measured partitions of absorption mechanisms. The data do not justify assuming that all severe CKD patients peak one hour after every cannabis product.

Did THC metabolites change in the same way?

No. Cannabinoid metabolism creates several analytes with different kinetics, analytical interpretations and potential relevance to effect. THC-OH is pharmacologically active. THC-COOH and its glucuronide are indicators of metabolism and disposition, not interchangeable with parent THC at cannabinoid receptors. The highest numerical metabolite median does not identify the compound responsible for dizziness or prove timing of use.

These comparisons demonstrate why a discussion that only says “metabolites increased” is scientifically incomplete. In particular, the stage 5 THC-COOH AUC did not reach the reported pairwise significance threshold versus controls, and the metabolite Tmax tests above were null. A report should identify the exact chemical species, transformation pathway, tested outcome and inference being offered.

Did the adjusted kidney-function analysis support an independent THC effect?

Not for THC AUC. In the original regression table, the unadjusted relation between creatinine-based eGFR and log-transformed THC AUC had coefficient −0.0067 and p<.001. After adjustment for age, body mass index and sex, the corresponding coefficient was −0.0031 with p=.217. Other eGFR specifications likewise lost significance after that adjustment. That result matters when anyone proposes using measured eGFR alone to calculate an individual THC concentration correction.

By contrast, adjusted associations for CBD and several metabolites remained significant. For creatinine-based eGFR and CBD AUC, the adjusted coefficient was −0.0089 with p=.005. Such adjusted observational associations still do not prove which physiological mechanism caused the difference. The authors also cautioned that the regression assumptions were not fully satisfied and that groups differed markedly in age and health. A clinical researcher might investigate altered enzyme function, protein binding, distribution or absorption; this dataset did not separately measure or establish those pathways.

What did the 24-hour urine collections reveal?

Total 24-hour urinary mass excretion of THC-COOH and THC-COOH-glucuronide was lower in the CKD groups than in controls. That is consistent with altered handling of measured metabolites, but lower urine recovery does not directly quantify liver production, parent THC clearance, or the fraction of dose absorbed. In reduced urine output, concentration and total excreted mass must be distinguished: a concentrated small volume and a dilute large volume can yield different concentration values but comparable excreted masses.

Parent THC and CBDA were not detected or were below quantification in the urine samples. THC-OH and THCA-A were quantifiable in only one urine sample apiece; CBD was quantifiable in ten urine samples and did not show a group difference. An assertion that all seven targets were reliably quantified in urine would therefore be false. None of these urinary measures substitutes for an observed parent-THC plasma concentration or a blood specimen contemporaneous with an alleged driving event.

Were participants more intoxicated or objectively impaired?

Participants rated ten symptoms on an unvalidated numeric scale from zero to ten. The group-mean dizziness score at 1.5 hours was approximately 0.1 in controls, 0.7 in stage 4 and 1.5 in stage 5, with both CKD groups significantly above controls at that assessment. These mean values are low on a ten-point scale and need not characterize every individual. Reported sensations of feeling high or intoxicated generally peaked around one to two hours. That item was added after the study began, limiting uniformity of ascertainment.

The study did not measure a standardized driving performance endpoint, psychomotor divided-attention task, field sobriety examination, crash risk, clinical impairment threshold or legally operative concentration. Dizziness is a subjective symptom, not a calibrated driving-fitness classifier. Group-level dizziness can arise through exposure and through nonpharmacologic causes, including other medications, fasting, blood pressure, baseline illness or expectancy. It is scientifically incorrect to translate a mean dizziness rating directly into proof that a named person was impaired during a particular incident.

Why do the same dose and higher concentration not prove a personal correction factor?

A dose–concentration relationship depends on route, actual bioavailability, volume of distribution, clearance, elapsed time and biological covariates. The two sprays contain a nominal product dose; the measured plasma profile reflects absorbed and distributed drug after this specific oromucosal administration. Investigators could not apportion swallowed versus mucosally absorbed cannabinoid. The different medians do not provide a validated multiplicative conversion applicable to a smoked product, an edible, or an individual’s historical blood sample.

Even within the sampled groups, concentration profiles varied widely. Stage 5 consisted of nine individuals, and its THC AUC minimum and maximum extended from 1.56 to 18.09 h·ng/mL, with major overlap with the other groups. It would be hazardous to use the ratio of group medians—4.31 divided by 2.76, approximately 1.56—as a retrospective multiplier for any patient with stage 5 CKD. Such a shortcut ignores matrix, dose, use history, sampling time, combined disease factors and uncertainty in the underlying group comparison.

What are the central study limitations?

The research enrolled a modest clinical sample after difficulty recruiting sufficient participants, particularly for stage 5. It was unblinded and nonrandomized, with much older, more medicated patients in CKD groups. It tested a single relatively low controlled oromucosal dose for 24 hours, not chronic daily dosing, accumulation to steady state or abrupt discontinuation. Cannabinoid metabolites remained measurable at the endpoint, so the study did not observe complete terminal elimination. Dialysis results were outside the report.

Laboratory assays measured total cannabinoid concentrations, not free unbound drug. Hydration status and body composition were not directly evaluated. The subjective symptom questionnaire was not validated and one item entered partway through enrollment. Several mechanistic explanations were discussed without directly measuring the mechanism in participants. Those limits do not erase the observed pharmacokinetic differences; they set the conditions under which extrapolation is defensible.

How should this evidence be used in forensic toxicology and litigation?

An accurate report can cite this experiment as evidence that severe CKD cohorts had statistically different plasma exposures after a controlled fixed Sativex dose. If a clinical record supplies eGFR, actual medication history and timed specimen results, the CKD finding may support a carefully bounded discussion of alternative pharmacokinetic explanations. It cannot itself establish the direction and magnitude of an unobserved case-specific THC concentration, the time cannabis was last consumed, the patient’s mental state, intentional behavior or accident causation.

For a defendant, witness, employee or patient with renal disease, the necessary audit begins with the actual specimen and test: was it venous whole blood, separated plasma, serum or urine; was parent THC or THC-COOH measured; was the result confirmed by an appropriately validated method; when was it collected relative to dosing or event; and what contemporaneous behavioral evidence exists? Without those facts, this study provides scientific context rather than an individual conclusion.

The proper evidentiary comparison also separates a concentration from an AUC, a nominal dose from absorbed drug, a metabolite from parent THC, and a questionnaire response from demonstrated functional incapacity. Renal function may change expected exposure without yielding a numerical excuse or a numerical presumption of impairment. A transparent expert can describe the distribution and adjusted null result without pretending there is a precise clinical formula where none was tested.

Frequently asked questions

Does stage 5 CKD always mean higher THC than stage 4 CKD?

No. Stage 5 had a slightly larger median THC AUC, but the CKD stage 4 versus stage 5 AUC comparison was not statistically significant. Both groups contained variable individuals, and the stage 5 sample comprised only nine nondialysis patients.

Did the paper directly measure blood THC?

It collected blood and then separated and analyzed plasma. It did not report an independently quantified forensic whole-blood parent THC series. Plasma numbers cannot be substituted silently for legal whole-blood results.

Could the higher THC AUC in CKD prove that impaired renal function alone caused it?

No. Group differences were observed, but age, body mass, medications and health also differed. After age, sex and BMI adjustment, the eGFR–THC AUC association was not statistically significant (p=.217 in the creatinine-eGFR model).

Did CBD exposure change with kidney disease?

Yes. Measured median plasma CBD AUC was higher in both CKD groups than controls, and the adjusted creatinine-eGFR association remained significant (p=.005). That still does not establish a patient-specific dose correction or prove an individual adverse effect.

Why were urine metabolite results different from plasma results?

Urinary total excretion and circulating plasma exposure describe different processes. Twenty-four-hour urinary THC-COOH and its glucuronide excretion were lower with CKD; that finding cannot be recast as lower absorbed THC dose or lower plasma parent THC.

Did participants with CKD fail a driving or sobriety test?

No such validated test was performed. Investigators collected subjective symptom scores and plasma drug measurements. More dizziness on average cannot establish that any identifiable driver was impaired or caused a collision.

Can these findings be applied to a dialysis patient or a chronic edible user?

Not directly. Dialysis participants were excluded from the reported analysis; the experiment evaluated one low fixed oromucosal dose under fasting conditions. Repeated dosing, products, routes and patient circumstances require separate evidence and validation.

Primary scientific reference

Sønderskov MB, Khatir DS, Kjærgaard KD, et al. Pharmacokinetics and Side Effects of Δ9-Tetrahydrocannabinol and Cannabidiol in Patients with Different Stages of CKD. Kidney International Reports. 2025;10(3):707–719. Original complete primary experimental publication: doi:10.1016/j.ekir.2024.12.030.