The 1998 report describes a short-term erectile-response signal in ten men with erectile dysfunction of no known organic cause. It compared Melanotan II with vehicle placebo using six-hour RigiScan monitoring. This is evidence about a small, selected crossover experiment, not proof of sustained treatment benefit or a basis for using a research material.[1][2]
A later report by the same investigators helps clarify the study methods, but combines this experiment with a different ten-person study. This record keeps those populations and results separate. The later publication is not an independent replication of the 1998 findings.[3][4]
Study at a glance
| Field | Study record |
|---|---|
| Original publication | Wessells et al., Journal of Urology, August 1998; 160(2):389–393 |
| Identifiers | PMID 9679884; DOI 10.1016/S0022-5347(01)62903-3 |
| Population | Ten men with psychogenic erectile dysfunction; all ten reportedly completed the experiment |
| Comparison | Melanotan II versus vehicle in a double-blind crossover |
| Measurement | Six-hour real-time RigiScan; presence, duration and rigidity of erections |
| Later clarification | The investigators' 2000 report distinguishes this psychogenic group as Study 1 from an organic-risk-factor group, Study 2 |
| Evidence boundary | Acute physiological response, not demonstrated durable function, successful intercourse or long-term net benefit |
Who was studied?
The original publisher's indexed results give a mean age of 47.4 years and a range of 22–59. Baseline nocturnal monitoring found erectile activity in all participants.[2] That selection matters: evidence in people with retained nocturnal erectile capacity cannot simply be generalized to every vascular, neurologic, hormonal, medication-related or mixed cause of erectile dysfunction.
Completion is also narrower than complete outcome reporting. Having no reported withdrawals does not establish that every planned measurement was captured, every comparison was reported, or all important harms were followed long enough. The available material does not permit reconstruction of each man's full sequence of measurements.
For a broader framework, see how to read a research study.
Crossover sessions were monitored at home
The investigators' 2000 account describes randomized order, two active and two vehicle sessions, and at least 48 hours between administrations. RigiScan monitoring took place at home, with instructions to remain awake and avoid erotic stimuli. These are later-reported methodological details, not a newly recovered original protocol.[4]
A crossover can reduce between-person variation because participants contribute observations under both conditions. It does not make treatment order, carryover or repeated-measure analysis irrelevant. A stated interval is not proof that all pharmacological or behavioral effects from one period disappeared before the next. We did not obtain a dated protocol, allocation list or pharmacokinetic justification that independently verifies that assumption.
The later discussion also acknowledges possible stimulation from the RigiScan device and uncertainty about adherence to the wakefulness instruction.[4] Accordingly, this was neither a continuously supervised laboratory experiment nor a study of unrestricted everyday sexual activity. Those are different settings, with different possible influences on the measurements.
The randomization, masking and control-group guide explains why a design label alone does not establish successful concealment or masking. Recognizable side effects could also reveal a condition; no formal masking-success assessment was available for this review.
What did the original report find?
The original abstract reports apparent erections in 8 of 10 participants after Melanotan II. Mean time with tip rigidity greater than 80% was 38.0 minutes with active treatment and 3.0 minutes with vehicle, with P=.0045.[1] These are distinct summaries: one counts participants with an apparent response; the other describes time above an instrument threshold.
The later report's Study 1 table reproduces the 38.0/3.0-minute comparison and gives mean tip-rigidity activity units of 78.4/9.8. Its table labels the rigidity band 80–100%, whereas the original abstract uses greater than 80%.[4] We retain those source-specific descriptions rather than claiming to have independently resolved how observations exactly on the boundary were classified.
Activity units are an instrument summary, not minutes, percentages of men, or a probability of successful intercourse. The ten participants are also not ten independent parallel-group participants on each side of the comparison.
Statistical significance does not settle clinical importance
The reported P value is evidence about a particular statistical comparison, conditional on its analysis assumptions. It is not the probability that the intervention works, a measure of effect size, or proof that an individual would benefit.
The later methods specify one-tailed Student's t tests for RigiScan comparisons and Fisher's exact tests for side effects. They do not clearly identify a paired test, a crossover model accounting for period and sequence, or multiplicity control.[4] This review therefore does not relabel P=.0045 as an independently verified two-sided, multiplicity-adjusted confirmatory result.
Individual paired differences, a confidence interval for the treatment contrast and a complete analysis dataset were not available. Subtracting the displayed means is not a substitute for estimating uncertainty using the actual repeated observations. Likewise, treating multiple administrations from one participant as unrelated people would use the wrong unit of analysis.
For interpretation, keep three questions apart:
- Did the measured contrast favor active treatment?
- How uncertain is that contrast under an appropriate analysis?
- Does the measured outcome establish an important and sustainable benefit?
The first does not answer the other two. See confidence intervals, P values and multiplicity and surrogate endpoints versus clinical outcomes.
Do not import the second study's desire results
The later paper's structured sexual-desire comparison and its lost RigiScan records belong to Study 2, not this original psychogenic cohort.[4] They must not be presented as additional outcomes or attrition from PMID 9679884.
The accessible original sources do not provide enough detail to establish a validated desire benefit, or a rigorously defined null desire result, in these ten men. Absence of a demonstrated benefit is different from demonstrating no effect. A rigidity result cannot substitute for measures of desire, satisfaction, distress, relationship experience or fertility.
The original indexed results describe no statistically significant change in total or free testosterone.[2] That observation does not establish hormonal equivalence or an endocrine indication. It should not be converted into a testosterone-enhancement claim.
Harms in ten participants are not a safety profile
The original abstract reports more nausea, stretching/yawning and appetite reduction with Melanotan II; no symptom required treatment in that experiment.[1] It does not supply a complete participant-by-symptom table, severity distribution or long-term follow-up.
The later report pools some safety results across both studies and uses administration-level denominators.[4] Those numbers are not ten-person risk estimates. We do not reconstruct original-cohort event rates from pooled totals or treat an apparently tolerated second exposure as evidence of long-term tolerability.
“Did not require treatment” does not mean harmless. The ability to detect a short-lived symptom is different from the ability to estimate uncommon, delayed or cumulative harm. Neither completion nor a favorable instrument result resolves that uncertainty. See adverse events, serious events and tolerability.
Current FDA concerns are separate evidence
On September 3, 2026, FDA's compounding-safety page placed Melanotan II in its withdrawn-nominations table, not its current Category 2 table. The row describes potential immunogenicity related to aggregation or peptide impurities and cites reports involving melanoma, posterior reversible encephalopathy syndrome, sympathomimetic toxidrome and priapism.[5]
These later reports are not events observed in the 1998 participants. Listing them does not establish an incidence rate or prove causation in every reported case. Equally, the small crossover cannot negate later safety concerns. Withdrawal of a nomination and an agency safety-risk entry are not a finished-product approval or a complete determination covering every legal or compounding circumstance.
The trial and FDA status guide explains how to keep such records separate.
Melanotan II is not bremelanotide or Vyleesi
The reported Melanotan II structure is an acetylated cyclic heptapeptide ending in an amide: Ac-Nle-c[Asp-His-D-Phe-Arg-Trp-Lys]-NH2.[2][4] Vyleesi's label describes bremelanotide acetate with a free acid at the peptide's carboxyl terminus, in a specific sterile finished formulation and autoinjector.[6]
An amide-versus-free-acid distinction is not erased by similar sequence names or shared receptor activity. Nor does a published chemical structure authenticate a separate lot. See peptide sequence and structural identity and active ingredients versus formulations.
FDA's application history records Vyleesi's original approval on June 21, 2019 and a labeling supplement on October 23, 2020. The current DailyMed page is updated November 13, 2025, while its prescribing text is revised March 2024. These are different dates and records, not interchangeable approval events.[6][7]
The labeled indication concerns acquired, generalized hypoactive sexual desire disorder in specified premenopausal women; the label excludes use for men or sexual-performance enhancement.[6] That product-specific approval does not transfer to Melanotan II, a compounded preparation or a catalog material. The approved-drug versus research-material guide develops that distinction; the melanocortin library organizes related but nonidentical entities.
Registration, funding and access limits
No registration identifier or dated original protocol/analysis plan was found in the sources reviewed. That is a bounded finding, not proof that no historical record exists anywhere or an allegation of misconduct. Without a contemporaneous plan, the original outcome hierarchy and analysis decisions cannot be independently certified. See registration, protocols and outcome switching.
PubMed indexes federal research support from NIDDK and NCI.[1] The later report separately acknowledges University of Arizona institutional support.[4] These statements do not establish who controlled every original design, analysis or publication decision, and should not be merged into an invented complete sponsor-role account.
For this update, the original abstract, official bibliographic record and indexed publisher excerpts were accessible; the original full-paper route returned an access error. The complete public author-uploaded text of the 2000 report was read, including its appendix, but it is a different publication. No complete original-paper PDF, original-table visual audit, separate protocol/SAP or participant-level reanalysis is claimed.
The September 3 PubMed record contained no listed correction/retraction relationship. That observation is limited to the record checked, not a guarantee that no notice exists elsewhere.
What this evidence can support
This is a small historical experiment with an acute physiological signal and material reporting limits. It does not establish durable or consistently repeatable benefit, improvement in all causes of erectile dysfunction, successful intercourse, long-term safety, or equivalence to an approved treatment.
Keeping these boundaries visible is more useful than turning a short monitored response into a general therapeutic or product-quality claim.
Frequently asked questions
How many men were in the original study?
Ten. The later combined report includes another ten-person study; its pooled population must not replace the original denominator.[1][4]
Was this a parallel-group trial?
No. It used active and vehicle conditions within participants. A valid analysis still needs to handle repeated observations appropriately.
Was monitoring performed in a laboratory?
The later investigators' report says monitoring occurred at home. It was a constrained six-hour session, not an unrestricted effectiveness study.[4]
Does the rigidity result prove successful intercourse?
No. An instrument threshold is not direct evidence of intercourse, satisfaction or durable function.
Did the original study establish increased sexual desire?
No established desire benefit can be assigned to this cohort from the reviewed sources. The structured desire comparison in the later report belongs to Study 2.[4]
Did completion establish safety?
No. A small completed experiment cannot quantify uncommon or delayed risks.
Does Vyleesi approval cover Melanotan II?
No. It concerns a different named active ingredient, finished formulation and labeled population.[6][7]
Primary and official sources
- Original study abstract and bibliographic record, PMID 9679884
- Original publisher record, DOI 10.1016/S0022-5347(01)62903-3; full-paper access limited
- Investigators' later 2000 report, publisher metadata and abstract
- Investigators' later report, public author-uploaded full text
- FDA, substance-specific compounding safety concerns
- DailyMed, Vyleesi label, updated November 13, 2025; text revised March 2024
- FDA, Vyleesi application 210557 and approval history
Review and corrections
Sources checked September 3, 2026 (America/Edmonton and UTC). Serum Research is responsible for this editorial summary and its corrections. No independent clinical, statistical, analytical or regulatory specialist review is claimed. This is nonmedical educational material, not treatment guidance, a product recommendation or a quality/equivalence conclusion.
Read how we review evidence, browse the study records and research library, or contact us about a correction.