On 2025-11-23 17:32:31, user Charlotte Strøm wrote:
In the following “text in italics – inside quotations marks are copy-pasted from the reference in question." Underlining and/or bolded text are done by me.
1. SPIN AND FRAMING<br />
The title of the preprint is: “Randomised trial of not providing booster diphtheria-tetanus-pertussis (DTP) vaccination after measles vaccination and child survival: A failed trial”<br />
1.1 Framing neutral findings as abnormal or disappointing<br />
The authors consistently imply the results are “unexpected” or “contradictory”, rather than acknowledging that the RCT failed to support earlier observational findings.
Examples<br />
“A failed trial” says the title ->The trial did not “fail”: it ran, randomised 6500+ participants, and produced valid estimates showing no harm of the DTP vaccine. Calling it “failed” is a framing tactic that positions the result as an error rather than what the data showed.
Page 8, lines 11-12: “The was no difference in non-accidental mortality … the HR being 0.84 (0.52–1.37).” ->This is an appropriate stating of results, but the subsequent framing undercuts it.
Page 8, lines 22-24: “Since no beneficial effect of not giving DTP4 was found, contradicting many observational studies… possible interactions were explored…” -> This subtly frames the RCT as problematic because it contradicts earlier observational research, rather than recognizing that RCTs supersede observational evidence.
Page 10, line 2:“The present RCT is therefore an outlier which needs an explanation.” -> This is spin: the RCT is not an "outlier" needing explanation; observational studies - upon which the research hypothesis are based - are know to have confounding and are biased. CONSORT encourages presenting results without exaggeration or defensive justification.
1.2 Causal interpretations of non-significant results<br />
The authors imply meaningful patterns where no statistically reliable findings exist.
Examples Page 8–9 (exploring interactions despite explicitly stating low power): “There was one significant interaction … DTP strain … observed only for females.” (p=0.05)
No correction for multiple testing; >20 interactions tested -> This is classic exploratory-analysis spin.
1.3. Hypothesis-confirming language<br />
The manuscript repeatedly positions NSE hypotheses as foundational truths rather than unproven claims.
Example Page 3, lines 5-7: “Several studies inidcated… beneficial non-specific effects… more pronounced in females.” -> These were observational or post-hoc analyses, being framed as established background biases the narrative.
1.4 Framing underpowering as the main explanation<br />
Repeated emphasis that the trial was “strongly underpowered” serves to discount the main finding.
Examples Page 8, lines 13-15: “...the trial was planned with 3% annual expected mortality rate… observed rate was 81% lower… we had 65% fewer deaths…” -> This is accurate but placed repeatedly tthroughout the text to frame the null result as flawed.
Page 9, lines 18-21: “The RCT was strongly underpowered… mortality declined …” -> The authors do not consider that a null finding study is plausible.
2. CONSORT NON-COMPLIANCE <br />
2.1. Missing or unclear prespecified primary outcome<br />
CONSORT requires explicitly stating primary and secondary outcomes and linking to a prespecified Statistical Analysis Plan (SAP).
Issues: The manuscript says: Page 5, lines 25-27: “The outcomes were all-cause non-accidental mortality and hospitalisation, as well as sex-difference…”<br />
-> It is unclear which of these is the primary outcome. Mortality? Hospitalisation? Sex-differential mortality? AND - there is ...
-> No link to protocol-defined hierarchy.
2.2. Discrepancies between protocolled numbers and intervention<br />
There are discrepancies between numbers stated in the publicly available protocol and study record at http://clinicaltrials.gov and the numbers appearing in the preprint. The intervention described in the preprint is not aligned with the protocol.
These discrepancies are unexplained in the preprint. The preprint states that DTP3 has been reported elsewhere, but the reference that is included in the preprint (2) does not report on mortality data, moreover it includes both DTP3 and DTP4. And these protocol deviations are inadequately accounted for in the preprint.
It remains therefore unexplained what the actual flow of study subjects were, and it remains unclear what the results are from the DTP3+OPV+MV versus OPV + MV only – as stated in the protocol.
2.3. Randomisation procedure is not sufficiently described<br />
CONSORT requires allocation concealment method and sequence generation details
Example Page 5, lines 12-14: “...randomisation lots were prepared by the trial supervisor… kept in envelopes… mother asked to draw envelope…” -> No description of safeguards (opaque, sealed, sequentially numbered). -> Allocation was not blinded, but CONSORT requires explicit reporting of potential bias. DTP3 is not mentioned in the trial flowchart - figure 1.<br />
2.4. Lack of intention-to-treat analysis <br />
CONSORT requires ITT or explanation for deviations.
Example Page 7, lines 1–3: “All children with follow-up and who received the per-protocol intervention were included in the analyses.”<br />
-> This is per-protocol only, inappropriate for a superiority RCT intended to detect harm.
-> No ITT analysis is presented.
2.5. No reporting of missing data handling<br />
CONSORT requires transparent handling of missing outcome data.
Example Page 7, line 14: “No imputation for missing data was done.” -> But the extent of missing data is not reported for mortality or hospitalization outcomes.
2.6. Discussion includes non-evidence-based explanations, violates CONSORT as Discussion should reflect results, not speculation<br />
The discussion drifts into immunological theory and historical interpretations unsupported by trial data.
Examples: Page 10, lines 4-6:“...likely that immune mediated NSEs are more pronounced when mortality is high…”<br />
Page 9-10 (multiple paragraphs): Repeatedly argues unexpected null results require explanation. -> This is speculative; not based on data reported from this RCT.
2.7. Lack of balanced discussion<br />
CONSORT item 22: discuss both limitations and strengths. -> The manuscript heavily emphasises limitations (underpowering, interventions, etc.), but does not discuss the strength of randomisation or lack of harmful signal which is odd considering the research hypothesis of the trial.
3. OVERALL REFLECTIONS ON THE IMPACT OF SPIN, FRAMING, AND CONSORT DEVIATIONS<br />
Altogether, it seems to be rather unusual for researchers to put the trial down already in the headline, downright devaluating the trial. The authors are known to advocate detrimental effects of the DTP vaccine, a hypothesis that is based on purely observational studies (3, 4), and very small numbers that have not managed to replicate even by the same group of researchers (5).
This preprint reports results from a large-scale randomized trial, outranking observational studies in the hierarchy of evidence. Hence – making use of “A failed trial” appears to be an attempt to frame the results as invalid, which is ethically disturbing and highly inappropriate towards trial participants and readers.
p. 10:“The present RCT is therefore an outlier which needs an explanation. The drop in power due to the declining mortality rate may not only have lowered the possibility of finding significant tendencies; it is also likely that the immune mediated NSEs are more pronounced when mortality is high, so when mortality declines by >80%, the residual deaths may be less likely to be affected by immunological changes.”<br />
There seems to be a deliberate misinterpretation, unsubstantiated, and highly speculative. It is difficult not to read this in any other way than as a deliberate attempt to spin the results, frame them to be perceived according to the authors’ hypothesis about DTP having detrimental effects and increasing child mortality. Spinning results is defined as questionable research practice (6). The study was a null finding study, not an outlier.
There were no signs of more pronounced negative NSE, i.e., higher mortality in the child participants, who got DTP with, or after the measles vaccine. However, the primary outcome analysis demonstrated that this trial is a null finding study and thus the hypothesis was rejected.
3.1.Spinning the facts around other interventions.<br />
Several times in the preprint, the authors argue that other health interventions affected the trial conduct and the results.
Examples<br />
Page 1; “During the trial period many new interventions, including many national health campaigns, were carried out.”<br />
and <br />
“due to the large number of health interventions, not envisioned at the initiation of the trial, a limited part of the follow-up was a comparison between DTP4+OPV4 vs OPV4 as the most recent vaccinations”<br />
Page 6:“Other interventions and interactions. As the number of routine vaccinations and national health campaigns vaccinations increased through the 1990s and the 2000s, it has become increasingly clear that there are numerous interactions between different health interventions, such as vaccines and micronutrient supplementation, which are usually not taken into consideration in planning a vaccination programme. For example, the sequence of vaccinations, the time difference between non-live and live vaccines, and booster exposure to the same vaccines all had impact on the mortality levels. In addition, most vaccines have sex-differential NSEs (16). Since children were enrolled at 18 months of age, there were numerous possibilities for interactions with (a) national health intervention campaigns before enrolment; (b) participation in previous RCTs; and (c) national health campaigns after enrolment in the trial.”
Page 10: “trials of NSEs were planned more or less as vaccine efficacy studies. However, it has become increasingly clear that there are interactions with other routine vaccinations, vaccination campaigns, and other interventions affecting the immune system like vitamin A (16,19,20). Hence, in the present RCT we examined possible interactions with campaigns before enrolment, previous RCTs, and campaigns given after enrolment.”
-> The reader is left with the impression that a series of other factors influenced the trial and possibly invalidated the results. However, this was a randomized trial set-up which to a great extent compensates for any potential confounding effects, ie. other interventions that may have affected the outcome; but they will do so in both the intervention and comparator group.
Moreover, from table 1 of the preprint – Baseline characteristics – it would seem that the authors tend to put too much weight on multiple other factors as the trial appears to be well randomized.
Finally, if it in fact was true that this trial was influenced by other RCTs, health interventions, or campaigns, then this argument applies to all trial data originated from this research group in Guinea Bissau and consequently invalidates all of them.<br />
Again it is remarkable that the authors put down their own trial, spin the data and frame them into letting the reader believe that the trial is worth nothing at all. This is not in accordance with appropriate reporting standards as per CONSORT (7).
3.2. Spinning the facts around the succession of vaccines<br />
p.3 “high-titre-measles-vaccine (HTMV) was protective against measles infection, but surprisingly, it was associated with higher female mortality, when tested against STMV (5,6). Hence, NSEs could be beneficial or deleterious and they were often sex-differential.
References 5 and 6 are self-citations and based on post hoc re-analyses. The hypothesis that the DTP – following HTMV induced higher mortality remain highly speculative and never replicated. A more likely explanation would be that the HTMV was dosed too high resulting in measles infections, attenuated but still, which unfortunately in some cases increased the subsequent risk of mortality. This is notably a specific effect of the vaccine. However, as the authors advocate that the live (attenuated) vaccines are inferring beneficial effects and the non-live vaccines infer detrimental effects, a post-hoc narrative was constructed on the succession of vaccines having relevance. Importantly, this current preprint where the DTP vaccine is given alongside or not a live attenuated vaccine does not support this highly speculative hypothesis. On the contrary: if anything the results pointed towards DTP increasing child survival.
- OVERALL REFLECTIONS ON ETHICS
4.1. Troubling lack of ethical standards and compliance
p. 7 it is stated that the study was explained to mothers in the following way:
“...though DTP is highly protective against whooping cough, it can occasionally give adverse reactions or limit the effect of measles vaccine….”
This speculative hypothesis seems to be introduced in the study participant / guardian information material, although this was never defined as a research question in the protocol.
Moreover, the protocol states:
“Hypothesis: Not providing DTP together with or after MV is associated with a 35 % reduction in overall mortality and 23% reduction in hospitalizations.
Taking one step back – and reflecting just for a minute – it appears to be the wildest research question ever. How did the Ethics Committee and the relevant authorities allow for this largescale trial to be conducted in the first place? What could possibly justify a RCT of this magnitude based on an outrageous research question like the one that was raised in the protocol: A 35% reduction in mortality is expected from omission of a single shot of vaccine?
4.2. Underpowered or not<br />
The preprint states that the trial was “highly underpowered,” although 109% of the planned study population was enrolled. There seems to be a large contrast between how this trial and a recently reported trial (8) are interpreted based on whether there was a significant finding or not. These discrepancies indeed appear as tendentious framing.
A direct comparison of these two large RCTs conducted by the same research group – with vast discrepancies in the results (enrolment and conduct) as well as interpretation is available at this link: https://www.linkedin.com/pulse/review-preprint-reports-dtp-trial-nct00244673-charlotte-str%25C3%25B8m-awgtf/ <br />
4.3. Self-citation rate of 95%<br />
Nineteen of 20 references include members of the same author group – and are thus self-citations. This may reflect a general lack outside this group of scientific support to the NSE hypothesis and / or selective citation which is considered to be questionable research practice (6). A rule of thumb is that a self-citation rate above 15% raises suspicion of selective citation.
4.4. Reflections on the “Postscript” of the preprint<br />
It is truly a good thing that these results have finally come to light. The study subjects, their families, and the scientific community have been waiting for these data to be published.
The preprint is concluded by a lengthy postscript explaining the unusual long delay (14 years) in publishing the results from this trial.
"Postscript. We apologise for the late reporting. The implementation of the trial went quite different from the scheduled plans. In this older age group, more children than expected were registered by an ID and address that could not be followed. Funding was lacking for the PhD student to complete the data cleaning and analysis. Before funding could be obtained, the Guinean field supervisor had died which made it difficult to resolve some inconsistencies in data. The senior authors had too many other commitments. Finally, from 2020, the COVID-19 pandemic changed all priorities"<br />
These explanations may very well be seen as a result of hypocrisy, as members of this group of authors have published numerous papers – including reporting of several clinical trials during the past 14 years. Moreover during this delay it has been argued by members of the author group that an RCT with the exact same research hypothesis should be conducted (10):
“Almost 4 years after WHO reviewed the evidence for NSEs and recommended further research, IVIR-AC has now submitted for public comments two protocols of RCTs to measure the NSE impact of BCG and MV on child mortality:<br />
a. A BCG trial will compare mortality between 0 and 14 weeks of age for children randomized to BCG-at birth plus routine vaccines at 6–14 weeks of age vs. placebo at birth and routine vaccines at 6–14 weeks, with BCG at 14 weeks of age.<br />
b. An MV trial will compare mortality between 14 weeks and 2 years of age for children randomized to an additional dose of MV co-administered with DTP3 vs. placebo co-administered with DTP3.”<br />
According to http://clinicaltrial.gov the study hypothesis of NCT00244673.<br />
“DTP3/4+OPV+MV versus OPV+MV or DTP4+OPV4 versus OPV4”<br />
And even worse – it was claimed in the same publication Expert Review of Vaccines, Vol 17, 2018 – Issue 5 (10) that: "Science is also about accounting for all data. ... it has not been possible to conduct RCTs of DTP in high-mortality areas."<br />
There has evidently been a complete lack of willingness from the research group behind this trial to report on this null finding study that rejected the research hypothesis and rejected the hypothesis that the DTP vaccine has detrimental NSE. Such selection bias in reporting trial results on mortality is scientifically troubling and ethically both irresponsible and unacceptable.
References:
-
Agergaard JN, S.; Benn, C.S.; Aaby, P. Randomised trial of not providing booster diphtheria-tetanus-pertussis (DTP) vaccination after measles vaccination and child survival: A failed trial. In: Bandim Health Project IN, Apartado 861, Bissau, Guinea-Bissau; Department of Infectious Diseases, Aarhus University Hospital, Denmark; Bandim Health Project, OPEN, Department of Clinical Research, University of Southern Denmark/Odense University Hospital, Denmark; Danish Institute for Advanced Study (DIAS), University of Southern Denmark, Denmark, editor. 2025.
-
Agergaard J, Nante E, Poulstrup G, Nielsen J, Flanagan KL, Ostergaard L, et al. Diphtheria-tetanus-pertussis vaccine administered simultaneously with measles vaccine is associated with increased morbidity and poor growth in girls. A randomised trial from Guinea-Bissau. Vaccine. 2011;29(3):487-500.
-
Mogensen SW, Andersen A, Rodrigues A, Benn CS, Aaby P. The Introduction of Diphtheria-Tetanus-Pertussis and Oral Polio Vaccine Among Young Infants in an Urban African Community: A Natural Experiment. EBioMedicine. 2017;17:192-8.
-
Aaby P, Mogensen SW, Rodrigues A, Benn CS. Evidence of Increase in Mortality After the Introduction of Diphtheria-Tetanus-Pertussis Vaccine to Children Aged 6-35 Months in Guinea-Bissau: A Time for Reflection? Front Public Health. 2018;6:79.
-
Sørensen MK, Schaltz-Buchholzer F, Jensen AM, Nielsen S, Monteiro I, Aaby P, et al. Retesting the hypothesis that early Diphtheria-Tetanus-Pertussis vaccination increases female mortality: An observational study within a randomised trial. Vaccine. 2022;40(11):1606-16.
-
Bouter LM, Tijdink J, Axelsen N, Martinson BC, Ter Riet G. Ranking major and minor research misbehaviors: results from a survey among participants of four World Conferences on Research Integrity. Res Integr Peer Rev. 2016;1:17.
-
Hopewell S, Chan AW, Collins GS, Hrobjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 explanation and elaboration: updated guideline for reporting randomised trials. BMJ. 2025;389:e081124.
-
Thysen SM, da Silva Borges I, Martins J, Stjernholm AD, Hansen JS, da Silva LMV, et al. Can earlier BCG-Japan and OPV vaccination reduce early infant mortality? A cluster-randomised trial in Guinea-Bissau. BMJ Glob Health. 2024;9(2).
-
Benn CS. Non-specific effects of vaccines: The status and the future. Vaccine. 2025;51:126884.
-
Benn CS, Fisker AB, Rieckmann A, Jensen AKG, Aaby P. How to evaluate potential non-specific effects of vaccines: the quest for randomized trials or time for triangulation? Expert Rev Vaccines. 2018;17(5):411-20.