A diagnostic problem that looks deceptively simple
Excessive thirst and large volumes of dilute urine can point to very different underlying conditions.
In arginine vasopressin deficiency, the body cannot release enough of a hormone needed to conserve water.
In primary polydipsia, excessive fluid intake can produce similar symptoms without the same hormone deficiency.
Correctly distinguishing the conditions matters because treatment and monitoring differ.
A new study published in Nature Communications on 5 November 2026 examined whether an intravenous infusion of mannitol could help clinicians make that distinction by stimulating the release of copeptin, a measurable companion marker of vasopressin.
The research included a randomized crossover experiment in healthy adults and a small patient pilot.
It found a substantial hormone-marker response after mannitol in healthy volunteers and a measurable separation between the two patient groups.
The findings are encouraging but preliminary.
They establish biological plausibility and an early clinical signal, not a fully validated replacement for current diagnostic pathways.
Why copeptin is measured
Vasopressin, also called antidiuretic hormone, helps the kidneys retain water.
When vasopressin release is deficient, patients can lose large amounts of water through urine and experience persistent thirst.
Direct measurement of the hormone is difficult in routine clinical settings.
Copeptin is released alongside vasopressin and is more practical to measure, making it useful as a surrogate marker.
The difficulty is that a baseline copeptin result may not adequately distinguish the causes of excessive urination and drinking.
Clinicians therefore sometimes stimulate hormone release and measure the response.
Hypertonic saline–stimulated copeptin is regarded as an accurate approach, but it involves raising blood sodium under careful supervision.
That demands intensive monitoring and appropriate facilities.
The authors asked whether a different physiological stimulus might provide a workable route to diagnostic separation.
Mannitol was of interest because it changes osmotic conditions and can stimulate the vasopressin system.
The randomized crossover experiment
The first part of the study was a randomized, double-blind crossover trial involving 22 healthy adults.
Participants received intravenous mannitol and placebo in a randomized order, allowing investigators to compare each person’s biological response under two conditions.
Twenty-one participants contributed to the primary analysis. A crossover design can reduce some between-person variation because each participant acts as their own comparison.
Randomization helps address order effects, and blinding reduces some opportunities for expectation to influence measurement or interpretation.
Even so, the design was a proof-of-concept test in healthy adults rather than a clinical diagnostic validation study.
Healthy volunteers do not necessarily respond in the same way as patients with hormone deficiency, kidney disease, complex medication regimens or other health conditions.
The main endpoint was copeptin after mannitol infusion.
The investigators also needed to consider whether the physiological stimulus could be administered acceptably and safely under the study conditions.
What the healthy-volunteer results showed
At 90 minutes, copeptin after mannitol was estimated to be 2.29 times the level after placebo at mean baseline copeptin levels.
The reported absolute difference was 6.4 picomoles per litre, with a 95% confidence interval from 5.2 to 7.7 and a p-value below 0.001.
This was evidence of a strong measurable response to the stimulus in the controlled experiment.
It was not a measurement of diagnostic accuracy, because the participants in this arm were healthy volunteers rather than people whose diagnoses needed to be distinguished.
Nor does a strong average response guarantee that every person will respond in a clinically interpretable way.
The result supports the proposed biological mechanism: mannitol can provoke a measurable change in copeptin.
The next question is whether that response differs sufficiently and consistently between patient groups to support decisions about diagnosis.
A small patient pilot
The second part of the research enrolled 20 patients: 10 with arginine vasopressin deficiency and 10 with primary polydipsia.
All received mannitol in a single-arm pilot study.
Unlike the healthy-volunteer crossover experiment, this patient comparison did not randomize participants between mannitol and placebo.
Its purpose was to examine the separation in copeptin responses between the two established diagnostic groups.
After 30 minutes, the researchers reported a difference of 11.7 picomoles per litre between the groups, with higher copeptin in primary polydipsia.
The 95% confidence interval ranged from 6.5 to 17.0 picomoles per litre.
This difference is clinically interesting because patients with vasopressin deficiency would be expected to show a weaker hormone response.
But a difference between group averages is not the same as a reliable classification rule for an individual patient.
The number of participants was too small to settle an optimal cut-off or define diagnostic error rates with high precision.
What a useful diagnostic test must establish
Before a new stimulation test can be used widely, researchers need to know more than whether the average response differs between groups.
Clinicians need thresholds that identify disease with acceptable sensitivity and specificity.
They also need evidence about borderline cases, repeat testing, coexisting illnesses, medication effects and the consequences of false-positive or false-negative results. A false diagnosis of vasopressin deficiency could lead to unnecessary treatment and inappropriate fluid management.
Missing a true deficiency could leave a patient exposed to ongoing dehydration and other complications. A test also needs to work in the types of patients who actually present for evaluation, not only in carefully selected research participants.
The mannitol pilot points toward a potentially useful signal but does not yet establish a universally applicable decision threshold.
The authors explicitly call for larger confirmatory studies to validate diagnostic cut-offs and patient safety.
Why simpler testing would matter
A diagnostic procedure that requires less intensive monitoring could potentially broaden access to accurate endocrine assessment.
In hospitals and specialist clinics, staff time, equipment and supervision requirements influence how readily a test can be offered.
Patients may also experience practical burdens from long appointments or complicated preparation.
However, it would be premature to assume that a mannitol-based test is automatically simpler or safer in every setting.
Intravenous mannitol has physiological effects and must be considered in the context of a patient’s hydration, renal function and overall health.
The published experiment is not an instruction for patients to use mannitol themselves, and it does not support self-testing.
Any future clinical protocol would need carefully defined eligibility criteria, monitoring arrangements and stopping rules.
The potential operational advantage is a hypothesis to be evaluated alongside diagnostic performance, rather than an outcome already proven by the small pilot.
Understanding the statistical evidence
The healthy-volunteer trial reported a highly statistically significant copeptin difference between mannitol and placebo.
That finding supports the conclusion that the infusion stimulated the measured hormone marker under the experimental conditions.
The patient pilot, meanwhile, reported a positive group difference with a confidence interval that did not include zero.
These results should be interpreted at their respective levels.
The first concerns physiological stimulation; the second concerns an initial separation between diagnostic groups.
Neither alone tells us how often a future test would classify a patient correctly in a busy clinic.
Sample sizes of 21 in the main healthy-volunteer analysis and 20 in the patient pilot also limit the ability to detect uncommon adverse events or explore differences among subgroups.
Confidence intervals quantify uncertainty around estimated differences, but they cannot account for every source of selection bias or variation in future patient populations.
Why the two conditions are easy to confuse
The shared symptoms of excessive drinking and urination arise from different mechanisms.
In vasopressin deficiency, the kidneys receive too little hormonal signal to conserve water.
In primary polydipsia, sustained fluid intake can suppress the hormone system and produce dilute urine.
Clinical history, blood chemistry and other tests remain important, but they may not always resolve ambiguous cases.
A stimulation test seeks to reveal how the body’s regulatory system responds when challenged.
This is why copeptin has attracted diagnostic interest.
It can provide an indirect view of hormone release that a resting measurement may miss.
The present study fits within that broader effort to refine differential diagnosis.
Its contribution is the evaluation of mannitol as a possible stimulus, not the discovery of a new disease or a claim that symptoms alone can distinguish the conditions.
What would strengthen the evidence
A larger study should enroll patients across multiple centres and include a broad range of clinical presentations.
Investigators would need to predefine a candidate copeptin threshold, compare results against an appropriate diagnostic reference standard and report sensitivity, specificity and uncertainty.
Reproducibility would matter: a test that performs well once but varies substantially between visits may be difficult to use.
Safety monitoring should cover relevant physiological changes and adverse events, including outcomes in patients with conditions that might alter the response.
Researchers should also examine the time course of copeptin measurements to determine whether an earlier sampling point offers practical advantages without sacrificing accuracy.
Finally, an implementation study could compare staffing, cost, patient experience and workflow against existing tests.
Only after such evidence would it be reasonable to make strong claims about clinical substitution.
The takeaway for patients and clinicians
The study offers a credible early step toward a potentially more accessible method of distinguishing vasopressin deficiency from primary polydipsia.
In a controlled crossover experiment, mannitol stimulated copeptin release in healthy adults.
In a separate 20-patient pilot, the response differed between the two diagnostic groups.
Both findings are meaningful within the scope of the research, but they do not establish an approved diagnostic replacement.
The correct interpretation is that mannitol-stimulated copeptin warrants further investigation.
Patients experiencing persistent thirst or excessive urination should seek clinical evaluation rather than attempt to interpret the findings as a self-diagnostic guide.
The scientific value lies in identifying a promising physiological signal and defining the next validation steps needed before that signal can guide routine care.
For other medical research, see Research Today’s reporting on blood-pressure care and blood biomarkers in adolescent athletes.
Source Information
Study title: Mannitol-stimulated Copeptin for diagnosing Arginine Vasopressin Deficiency: a randomized controlled proof of concept and a single-arm pilot study
Authors: Sven Lustenberger, Friederike Riehle, Joseph G. Verbalis, Deborah R. Vogt, Cemile Bathelt, Cihan Atila and Mirjam Christ-Crain
Journal: Nature Communications
Publication date: 5 November 2026
Study design: Randomized double-blind crossover proof-of-concept trial in healthy adults followed by a single-arm patient pilot
Sample or scope: 22 healthy adults randomized, 21 in the primary analysis; 10 patients with vasopressin deficiency and 10 with primary polydipsia
Key finding: Mannitol increased copeptin 2.29-fold versus placebo at 90 minutes in healthy volunteers; patient groups differed by 11.7 pmol/L at 30 minutes
