The 1992 Bes study examined delta sleep-inducing peptide (DSIP) in 16 people with chronic insomnia. Its abstract reports higher sleep efficiency and shorter sleep latency compared with glucose placebo, but the investigators considered the statistically significant findings weak and potentially influenced in part by an incidental placebo-group change. Perceived sleep quality did not show a reported improvement, and the authors judged major short-term therapeutic benefit unlikely.[1][2]
This is a review of the accessible primary abstract and bibliographic information, not a full-paper audit. The full methods, numerical results, participant flow and safety data were not available for this review. That limit matters when interpreting both the favorable findings and the authors’ cautious conclusion.
Study identity and design
| Feature | What the accessible abstract supports |
|---|---|
| Publication | Bes, Hofman, Schuur and Van Boxtel; Neuropsychobiology 1992;26(4):193–197; PMID 1299794; DOI 10.1159/000118919 |
| Participants | 16 people described as having chronic insomnia |
| Design description | Double-blind, matched-pairs, parallel groups |
| Comparison | Historical DSIP study material versus glucose solution placebo |
| Observation | Five laboratory nights: adaptation, baseline, then three comparison nights |
| Measures | Sleep structure, objective polysomnographic measures, perceived sleep quality and tiredness |
The abstract states that half the participants received DSIP and half placebo. Eight per group follows from that description; it is not an independently verified number of complete records in every analysis. The historical route is identified in the abstract, but this page does not reproduce an exposure amount, preparation method or administration schedule.
Matching, masking and randomization are not interchangeable
The accessible description does not state how allocation was generated or whether it was randomized. It also does not give the matching characteristics, allocation-concealment procedure or a test of masking success. The appropriate conclusion is that randomization is unverified here, not that the study has been proved nonrandomized.[1][2]
Matching and double-blinding address different design questions. Neither term alone establishes a random allocation process. Conversely, an omitted abstract detail does not demonstrate that a safeguard was absent from the full study. The randomization, masking and control-groups guide explains these distinctions.
The revision therefore does not reconstruct how individual members of matched pairs were assigned. The abstract’s stated design and half-and-half comparison are the supported level of detail.
Five nights do not create five independent samples
The first laboratory night was used for adaptation and the second for baseline measurements, followed by three comparison nights.[1][2] Those are repeated observations in the same small participant sample, not five separate groups of 16 people.
The abstract does not expose the statistical model, how matching or repeated nights were handled, or how incomplete recordings were treated. Without those details, this review cannot verify the analysis unit or reconstruct an effect estimate. Repeated measurements may be useful, but counting nights as additional independent participants would overstate the sample size.
Laboratory adaptation, expectations and ordinary variation are possible considerations when interpreting sleep measurements. They are not established explanations of this study’s findings. In particular, the authors’ placebo-change caution does not prove that adaptation continued or caused the observed contrast.
Which findings were between groups?
The abstract’s comparison labels should remain attached to each result.[1][2]
| Measure | Reported result | What it does not establish |
|---|---|---|
| Objective sleep efficiency | Favored DSIP compared with placebo | The numerical difference, its precision or its clinical importance |
| Objective sleep latency | Favored DSIP compared with placebo | How many minutes differed or whether participants experienced a meaningful benefit |
| One subjective tiredness measure | Decreased within the DSIP group | A verified active-versus-placebo tiredness effect |
| Subjective sleep quality and other measures | No change reported in the abstract | Equivalence, an exactly zero effect or exclusion of all clinically important differences |
A within-group change answers a different question from a treatment comparison. Improvement from one time point to another can occur without demonstrating that the active intervention outperformed the comparator. The abstract does not supply a between-group tiredness estimate that would resolve that question.
Objective measurements and perceived experience are also distinct. A favorable laboratory result does not automatically mean better perceived sleep or improved next-day functioning. The pattern reported here is mixed, rather than a uniform improvement across sleep-related outcomes.
Why the authors’ caution matters
The investigators described their significant findings as weak and suggested that an incidental change in the placebo group might explain part of them. They concluded that short-term DSIP treatment was unlikely to provide major therapeutic benefit for chronic insomnia.[1][2] That interpretation should remain as visible as the two favorable objective measures.
The caution is conditional. It does not prove that placebo behavior caused the entire difference, that DSIP’s effect was numerically zero or that every possible effect was small. The abstract provides no group means, between-group differences, variability estimates, confidence intervals or exact P values from which to quantify those claims.
The adjective “weak” is the authors’ interpretation, not a substitute for an effect-size estimate. Similarly, a lack of reported change in another measure is not an equivalence result. See what negative or null study results mean for the difference between an unconvincing finding and evidence that rules out an important effect.
Significance, precision and patient importance
The abstract discusses statistically significant findings but does not expose a prospective primary-outcome hierarchy, a multiplicity procedure, an analysis plan or a power calculation. There were several measures and repeated nights, but this review cannot classify the tests as a verified confirmatory family or determine whether all comparisons were reported.
This missing detail is an access limit, not proof of selective reporting or an improper analysis. The full methods and numerical results would be needed to judge those issues. The confidence intervals, P values and multiplicity guide provides the relevant reading framework.
No clinical-importance threshold is supplied in the abstract. Consequently, favorable directions for sleep efficiency and latency cannot be converted into a claim that participants obtained a meaningful improvement. Nor can the authors’ skeptical conclusion supply the absent confidence intervals.
Missing records and participant characteristics
The abstract names 16 participants but does not report endpoint-specific analysis counts, withdrawals, excluded nights, recording failures or missing-data rules. Eight per group is therefore the described allocation scale, not a confirmed final-analysis denominator.
An unmentioned withdrawal is not proof of no withdrawal. At the same time, unavailable flow information does not tell us that data were lost or that missingness biased the result. Both assertions would exceed the evidence obtained. Trial estimands and missing data explains why the outcome population and handling of missing observations need explicit description.
Detailed ages, diagnostic criteria, comorbidities and concurrent-treatment restrictions were not established from the accessible abstract. The historical “chronic insomnia” description does not by itself demonstrate applicability to every current insomnia population or to care outside a laboratory.
Safety and material identity remain unresolved
The accessible abstract provides no usable adverse-event counts, safety-monitoring results, vital-sign findings or discontinuation details. It cannot establish that adverse events were absent, that the intervention was well tolerated or that it had an acceptable benefit–risk balance.[1][2]
The small sample and brief laboratory observation also do not establish uncommon, delayed or long-term safety. This review does not assume an undocumented safety-surveillance design or calculate a harm-detection probability from the eight-per-group allocation.
Manufacturer, counterion, excipients, analytical characterization, lot and sterility information were not available in the obtained abstract. Those details may exist elsewhere; they are not proved absent from the full paper. Without them, this record cannot authenticate or establish equivalence for another preparation, supplier, route or blend. A shared DSIP name is not a product-quality or clinical-equivalence assessment.
Funding, publication dates and review limits
The publisher lists University of Amsterdam psychology/Psychonomics and Academic Medical Center affiliations. These identify the authors’ institutions, not the complete funding arrangements, sponsor role or conflicts of interest. A full funding and disclosure section was not obtained.[1]
The citation is for the 1992 journal issue; the publisher separately labels its online publication February 22, 2008. This does not make the historical experiment a 2008 trial.
During the present review, the Karger page exposed the abstract and metadata but indicated that the full content was PDF-only and not currently accessible. The direct PubMed attempt in the review history returned an empty result; official indexed abstract text was separately available. A DOI access error was not retried. These limitations are not described as a successful full-paper or live-notice review.
No original protocol, statistical analysis plan or registration history was obtained, and no exhaustive registration or correction/retraction search was performed. This page does not renew older “no registration” or “no notice” statements as current verified findings.
Bottom line
The study’s abstract reports selected favorable objective sleep results but no reported improvement in subjective sleep quality, with an explicit caution about the strength and possible origin of the observed contrasts. It does not establish a meaningful clinical benefit, equivalence to placebo, general safety or the validity of a modern material.
The most useful reading is neither a positive headline nor a declaration of zero effect. It is a small, short, incompletely accessible study whose numerical uncertainty and full analysis cannot be assessed here. Start with how to read a research study or browse the Study Records library for related evidence-reading context.
Frequently asked questions
How many people were studied?
The abstract describes 16 participants, half receiving DSIP and half placebo, implying eight per group. It does not verify eight complete observations per group for every endpoint or night.
Was assignment randomized?
That is not established by the accessible abstract. It describes double-blind matched-pairs parallel groups but does not give an allocation method. Missing that detail does not prove nonrandomization.
Which outcomes favored DSIP over placebo?
Sleep efficiency and sleep latency favored DSIP in the abstract. Numerical treatment differences, confidence intervals and clinical-importance thresholds were not available.
Did tiredness improve compared with placebo?
The abstract reports a decrease in one tiredness measure within the DSIP group, not a verified between-group tiredness effect. It reports no change in subjective sleep quality.
Did the investigators prove that DSIP had no effect?
No. They considered the findings weak, potentially partly influenced by a placebo-group change, and unlikely to represent major short-term therapeutic benefit. That is not a numerical zero-effect or equivalence demonstration.
Did this review establish safety?
No. Usable adverse-event and safety-monitoring data were not available in the abstract, and the small, brief study cannot establish broad or long-term safety.
Can the historical study validate another DSIP preparation?
No. The obtained evidence does not establish another material’s identity, formulation, purity, sterility or comparability. It is not a recommendation for practical use.
Primary and official sources
- Karger primary publication: accessible abstract and metadata
- PubMed PMID 1299794: official bibliographic record
- DOI 10.1159/000118919: publication identifier
Editorial evidence review: September 3, 2026. Accessible abstract and metadata reviewed; full paper, numerical tables, figures, protocol and safety dataset not reviewed. No independent specialist review is claimed. Research information only, not medical advice or practical-use instructions. Report a correction.