No, the Cochrane Collaboration Review Does Not Support HPV Vaccination: Majority Laundering + Unreported COIs = Bought Bias

Low-quality studies with insufficient follow-up and again gloss over high-risk HPV type replacement. Second biased report is DOA. COIs, too. Majority Laundering fools no one.

In November 2025, the Cochrane Collaboration published what it framed as a landmark review on the population-level effects of HPV vaccination. The authors concluded that HPV vaccines “probably reduce” cervical cancer incidence by nearly 80% and asserted that serious harms like POTS, infertility, and premature ovarian failure were unlikely. But the review itself reads less like an objective summary of evidence and more like a surgical exercise in narrative construction. The most damning fact? Despite acknowledging that the majority of included studies—especially those purporting cancer prevention—were at serious or critical risk of bias, Cochrane still upgraded its certainty ratings and issued sweeping public-health conclusions after cherry-picking studies to include.

They used what can only be called Majority Laundering - the conflation of the majority of studies with a consensus of studies - a cardinal sin in meta-analysis. The overcooked report in overinterpreted. Share w/anyone giddy online about it: It is bogus. Details below.

Worse, this is not the first time Cochrane has evaded the core evolutionary virology concern of HPV vaccination: type replacement. In 2018, it quietly dropped its protocol plan to examine lesions caused by non-vaccine HPV types. In 2025, it repeated the move, ignoring mounting evidence that the suppression of HPV 16/18 may unmask or ecologically promote non-vaccine high-risk types like HPV 31, 33, 45, 52, and 58. The phenomenon—discussed in mechanistic and surveillance studies—is again glossed over entirely. For a full breakdown, see: HPV Type Replacement is Real – CDC’s Silence is Not.

Among the most egregious sleights of hand:

• GRADE Inflation: Cochrane elevates “moderate certainty” conclusions from observational studies where 9 of 20 cancer studies were at critical risk of bias, and only a handful included proper confounder adjustment.

• Narrative Data Carving: The authors split one study (Kjaer 2021) into three age-stratified effects, retained only the 0.14 IRR for those vaccinated under 16, and omitted the concerning IRR>1 in women vaccinated between 20–30.

• Safety Gaslighting: Though the review admits it found no community-level studies on serious adverse events, it still concludes with “moderate certainty” that harms like paralysis, GBS, and ovarian failure are not increased.

• Repackaging Registries: Instead of requiring lesion genotype data, the review aggregates ungenotyped CIN2+/CIN3+ registries—precisely the kind CDC uses to make claims of success—despite their inability to detect type replacement.

The result is not a neutral synthesis of global evidence but a directed public-relations product masquerading as scientific review. This isn’t an epistemic compromise—it’s a methodological surrender.

Majority Laundering: How Cochrane Manufactured Consensus

Throughout its 2025 HPV vaccination review, Cochrane engages in a pervasive and deliberate pattern of majority laundering—a rhetorical technique wherein mixed or divided evidence is reframed as broad consensus. This is achieved through two mutually reinforcing strategies:

1. Bias Suppression: Citing a ‘majority’ of studies that support a claim while downplaying or removing high-quality studies that do not.

2. Estimate Cherry-Picking: Emphasizing the most favorable relative risk (RR), incidence rate ratio (IRR), or odds ratio (OR) values while excluding, qualifying, or ignoring those suggesting null or adverse effects—even when they emerge from the same studies.

The result is not merely a distortion—it is the substitution of curated narrative for actual evidence.

Conflicts of Interest

Conflicts of Interest in the Cochrane HPV Vaccine Reviews: A Detailed Ledger

This section compiles and classifies all known and inferred conflicts of interest (COIs) for individuals and institutional contributors to the 2025 Cochrane real-world human papillomavirus (HPV) vaccine review, the preceding 2018 Cochrane prophylactic trial review, and the broader network of evidence production and guideline shaping. These include direct manufacturer (Merck & Co., Inc., also known as MSD outside the U.S. and Canada) payments, Merck-supplied products for studies, participation in Merck-funded advocacy organizations, and indirect financial or reputational entanglements.

---

I. Severe Conflicts of Interest: Direct Ties to Merck/MSD (Financial or Product Provision)

Dr. Maria Kyrgiou
- Co-author of the 2025 Cochrane population-level review; co-author of Cochrane’s network meta-analysis (NMA) on HPV vaccines; principal investigator (PI) of the NOVEL trial.
- Declared in Cochrane: Gardasil 9 (Merck) provided for her NOVEL trial by MSD; authorship of the European Society of Gynaecological Oncology (ESGO) and European Federation for Colposcopy (EFC) HPV vaccination position paper.
- Not fully declared: The ESGO–EFC position paper (2019) discloses institutional research and educational support from MSD, personal support from MSD for travel and speaking engagements. Also supported by Hologic.
- Unreported in Cochrane: These financial relationships with Merck.

Dr. Marc Arbyn
- Lead author of the 2018 Cochrane HPV trial review; ESGO/EFC position paper author; Belgian national screening advisor; member of the HPV Prevention and Control Board.
- Declared in 2018 Cochrane: Received travel grants from MSD–Sanofi Pasteur and GlaxoSmithKline (GSK) (ceased 2008); institute received grant from Sanofi–Pasteur–MSD for the Surveillance of the Effects of HPV Immunization in Belgium (SEHIB) study.
- Not declared in Cochrane: - Received approximately €48,000 from MSD–Sanofi Pasteur for SEHIB. - Former GSK HPV advisory board member. - Programme committee member of EUROGIN (European Research Organization on Genital Infection and Neoplasia), a conference where Merck has platinum sponsorship. - Active participant in the Merck-funded HPV Prevention and Control Board.

Dr. Lauri E. Markowitz
- Lead epidemiologist for HPV vaccine implementation at the U.S. Centers for Disease Control and Prevention (CDC); co-author of 2018 Cochrane HPV trial review; advisor to World Health Organization’s Strategic Advisory Group of Experts (WHO SAGE).
- Declared in Cochrane: Treated as conflict-free based on public sector employment.
- Not declared in Cochrane: - Co-authored a Medscape continuing medical education (CME) module titled “HPV Vaccination: Real-World Data,” funded by an independent educational grant from Merck.
- Contributor to CDC’s Morbidity and Mortality Weekly Report (MMWR) recommendations endorsing Gardasil 9.

HPV Prevention and Control Board (Institutional)
- Board cited in multiple Cochrane references and discussion points.
- Not disclosed: The Board is hosted by the Vaccine and Infectious Disease Institute (University of Antwerp) and funded by unrestricted grants from Merck, GSK, Abbott, and Sanofi Pasteur.
- Outputs: Publications in Papillomavirus Research, BMC Proceedings, and Preventive Medicine Reports that support HPV vaccine expansion and safety.
- Board members with Merck ties: Includes Arbyn, Francisco Xavier Bosch, and other figures prominent in HPV policy.

ESGO/EFC Position Paper Group
- Cited throughout Cochrane as an expert consensus.
- Funding disclosures (not acknowledged by Cochrane): - Maria Kyrgiou – institutional and personal MSD support. - Elmar Joura – MSD travel and speaking support; principal investigator for Gardasil trials. - Francisco Xavier Bosch – MSD and GSK institutional support. - Murat Gultekin – MSD personal travel and speaking support.
- These disclosures appear in both the 2019 HPV vaccination and 2020 cervical screening ESGO/EFC statements.

Gardasil-9 Randomized Controlled Trials (RCTs)
- Core evidence base across all Cochrane HPV reviews.
- Disclosed in 2025 Cochrane NMA: 44 RCTs analyzed; 18 funded by Merck, 25 by GSK, and others by Sanofi Pasteur and Innovax.
- Not emphasized: The meta-analyses that underpin Cochrane’s claims of efficacy rely overwhelmingly on manufacturer-run or manufacturer-funded trials.

---

II. Medium-Level Conflicts of Interest: Pharma-Funded Activities in Related Domains

Dr. Emma J. Crosbie
- Co-author of the 2025 Cochrane real-world review.
- Declared: National Institute for Health Research (NIHR) funding; honoraria from GSK and Astellas; research grants from Roche and Novosanis.
- Context: Leads Peaches Womb Cancer Trust, which has participated in regulatory evaluations of Merck’s pembrolizumab–lenvatinib (Keytruda combination). No documented direct Merck funding.

Self-Controlled Case Series (SCCS) Study Authors
- Authors of adverse event safety studies included in Cochrane’s synthesis.
- Not disclosed: Many of these studies are funded by CDC or national public health authorities who work closely with Merck during vaccine rollout. The COIs of contributing surveillance systems are not presented.

WHO SAGE–Cochrane Response Relationship
- WHO commissioned Cochrane Response (led by Henschke and Bergman) to assess one-dose HPV vaccine evidence.
- WHO’s deliberations rely heavily on ESGO, HPV Prevention and Control Board, and CDC outputs.
- Authors declared “no conflict” in WHO documents, despite operating within a heavily Merck-influenced knowledge ecosystem.

---

III. Structural and Narrative-Level Conflicts (Not Attributed to Individuals)

Cochrane never discloses that many of the “expert consensus” documents it uses to dismiss type replacement concerns are produced by Merck-funded panels.

Cochrane’s own COI screening system failed to flag key role conflicts. For example, Jo Morrison—senior Cochrane editor and 2025 review author—served simultaneously as a UK NHS screening guideline lead while overseeing HPV vaccine evidence review, introducing systemic bias not related to funding but to role entanglement.

EUROGIN and its programmatic outputs, heavily cited in the HPV literature and by Cochrane-linked figures (Arbyn, Joura), are supported by Merck at a platinum sponsorship level. These sponsorships and repeated speaking roles by authors such as Joura are not flagged anywhere in Cochrane’s summaries or disclosures.

Surveillance data used in Cochrane 2025—such as HPV-IMPACT from the U.S. and registry-linked national studies in Denmark and the UK—derive from systems where Merck either co-developed surveillance architecture, provided grant support, or participated in technical working groups. The structural conflict from reliance on these surveillance platforms is not addressed.

Key safety studies synthesized in the review, particularly the SCCS designs, are authored by individuals (e.g., CDC, NHS, WHO vaccine advisors) with overlapping responsibilities for vaccine policy success. This dual role as promoter and evidence generator is never disclosed.

IV. Origin of Initial Cochrane HPV Work: CDC-Commissioned Analysis

The 2018 Cochrane prophylactic HPV trial review (CD009069) was developed under a collaboration that included authors working at the U.S. CDC (e.g., Lauri Markowitz), with downstream MMWR policy output directly referencing Cochrane synthesis.

While Cochrane itself did not accept direct CDC grant funding for the review, key contributors operated under CDC support structures, and CDC-endorsed trial evidence (with known Merck involvement) was prioritized.

---

V. Recommendations for Disclosure Reform

1. Cochrane must revise its editorial COI policy to include dual-role conflicts, including those arising from authors serving in simultaneous guideline/policy/industry advisory roles.

2. All institutional citations (e.g., HPV Prevention and Control Board, ESGO/EFC position papers, EUROGIN proceedings) must include a financial origin footnote.

3. Trial- and surveillance-level conflicts should be disclosed in the summary-of-findings tables, not buried in methodological appendices.

4. Policy-embedded evidence generators (e.g., authors of SCCS safety studies from vaccine-implementing bodies) should be tagged with functional-role conflicts to prevent narrative laundering of institutional outputs.

This ledger is not an indictment of individual bad actors—it is a structural diagnosis. A supposedly neutral, evidence-based summary cannot credibly declare “moderate certainty” while drawing upon a research ecosystem so extensively funded, shaped, or narrated by the manufacturer whose product is under review.

This is not transparency. It is institutional laundering.

Examples of Bought Bias in Action

1. Bias Suppression by Selective Counting

Cochrane’s treatment of cervical cancer studies sets the tone: out of 20 studies on risk, 9 were rated at critical risk of bias, 7 at serious risk, and only 4 at moderate or better. Rather than dealing with this imbalance, Cochrane simply excluded most of the dataset from the meta-analysis, pooling only five adjusted cohort studies to generate its 80% cancer reduction claim. That exclusion was not caveated; it was vanished.

Fully 16 of the 20 cervical cancer studies—80% of the total—were rated at serious or critical risk of bias and excluded from synthesis. This uncomfortable majority was stripped of interpretive weight and erased from all summary language.

The same approach was applied to CIN3+ incidence. Among 23 studies, 10 were rated at critical risk of bias, 12 at serious risk, and only one at moderate. Yet Cochrane declared that HPV vaccination “probably reduces” CIN3+ lesions, sweeping away the overwhelming burden of bias as if it never existed.

When it comes to adverse events—CFS/ME, POTS, Guillain-Barré syndrome, infertility, CRPS—the pattern is identical. For example, the CFS/ME evidence base consisted of one critical-risk study, three serious, two moderate, and three self-controlled case series (SCCS) rated low risk. Despite the nearly even split between concerning and null findings, Cochrane asserts no increased risk based on the SCCS data alone. Similarly, for POTS, only three studies existed—one serious, one moderate, and one SCCS—yet the conclusion remains an unqualified “no increased risk.”

Guillain-Barré syndrome was covered by 13 studies: five critical, seven serious, one moderate. Despite this 12-of-13 high-bias majority, Cochrane labeled the conclusion as “low-certainty evidence” of no risk. For CRPS, all three included studies were at serious risk, yet Cochrane confidently reported safety based on a single SCCS. Two studies on infertility were both rated serious, but that didn’t stop the authors from asserting a reassuring conclusion.

In the case of prevalent HPV infection—one of the most central outcomes—80 studies were reviewed. Of these, 31 were rated at critical risk of bias, 46 at serious risk, and just three at moderate. Still, Cochrane produced strong conclusions of vaccine effectiveness without any visible discomfort about the fact that over 97% of the evidence was methodologically weak.

This pattern extended to behavioral and reproductive outcomes. On sexual activity, six studies were included: two at critical risk, four at serious. For pregnancy and neonatal outcomes, eight studies: one critical, seven serious. Even on all-cause mortality—perhaps the most severe endpoint considered—only two studies were included, both at serious or critical risk. Yet all of these were synthesized into “no increased risk” narratives based on the thinnest slices of data.

We have to take their word on the risk of bias of the studies included and excluded because the inclusion/exclusion of studies in meta-analyses can be driver of the conclusions.

2. Estimate Cherry-Picking and Effect Size Distortion

The most illustrative example of cherry-picking comes from Kjaer 2021, a robust Danish national registry study that reported cervical cancer incidence rate ratios stratified by age at vaccination. The IRR for those vaccinated at age ≤16 was 0.14—an impressively low number. For the 17–19 group, it was 0.32. And for women vaccinated between 20–30 years old, it was 1.15 with a confidence interval of 0.88–1.50, suggesting possible lack of benefit or even transient harm. Cochrane incorporated the 0.14 into its pooled estimate, highlighted the 0.32 as supportive, and excluded the 1.15 from all meaningful narrative or synthesis. This is selective invisibility: the higher risk figure appeared only in a table, without comment or context.

The same kind of omission appears throughout the CIN2+/CIN3+ outcome literature. Several studies report RRs ranging from 0.7 to 1.1, especially in older or catch-up cohorts. Some estimates cross the null. Rather than grappling with this heterogeneity, Cochrane chose to synthesize only the most favorable values—such as 0.27 and 0.31 from Lei 2020 and Falcaro 2021—presenting these as if they were universally representative.

On safety, the effect estimates are often even more contradictory. For POTS, one case-control study reported an OR of 1.44 with a CI that includes increased risk. Another study returned 0.93. The SCCS estimate was 0.89. Cochrane picked the SCCS and declared “no association.” CFS/ME estimates were similarly inconsistent across studies, but only the SCCS were used to make the narrative whole.

In the case of Guillain-Barré syndrome, estimates such as OR = 1.12 and IRR = 1.26 were either ignored or brushed aside, despite being included in the dataset. Cochrane focused instead on the SCCS numbers (typically near 1.0) to dismiss concern.

When it came to non-vaccine high-risk HPV infections (types 31, 33, 45, 52, 58), several studies reported ORs above 1.0. One even reached 1.34. Other prevalence studies showed numerical increases. Rather than engaging with these signals as potential early indicators of type replacement, Cochrane concluded, “no convincing evidence” and avoided meta-analysis of these values entirely.

Related:

Ovarian failure and infertility studies also produced risk estimates in the 1.08–1.30 range. Though confidence intervals were wide and statistical significance inconsistent, the direction of effect was clear. Yet again, Cochrane defaulted to “no association.” Similar patterns were observed in pregnancy outcomes and mortality, where estimates like OR = 1.12 or IRR = 1.18 were stripped of interpretive significance due to p-values hovering above 0.05.

Conclusion: When Dissent Is Systematically Erased, Certainty Is Manufactured

Cochrane’s review does not transparently weigh the evidence. It filters, curates, and recodes discordant data into agreement—by dropping high-bias studies, spotlighting only favorable subgroups, and assigning certainty labels untethered from the full range of effect estimates.

This is not evidence synthesis. It is statistical manipulation in service of narrative control. When 6 out of 14 studies contradict your claim, you cannot claim majority support.

When RR >1.0 appears in your own cited studies, you cannot say “no risk” with a straight face.

Yet Cochrane did exactly that—over and over again.

This is the anatomy of manufactured consensus. And it cannot stand unchallenged.

Citations

Cervical Cancer – Age-Stratified IRRs

1. Kjaer SK, Dehlendorff C, Belmonte F, Baandrup L. Real-world effectiveness of human papillomavirus vaccination against cervical cancer. J Natl Cancer Inst. 2021;113(10):1329–1335. doi:10.1093/jnci/djab075

2. Lei J, Ploner A, Elfström KM, et al. HPV vaccination and the risk of invasive cervical cancer. N Engl J Med. 2020;383(14):1340–1348. doi:10.1056/NEJMoa1917338

3. Falcaro M, Castañón A, Ndlela B, et al. The effects of the national HPV vaccination programme in England, UK, on cervical cancer and grade 3 cervical intraepithelial neoplasia incidence: a register-based observational study. Lancet. 2021;398(10316):2084–2092. doi:10.1016/S0140-6736(21)02178-4

4. Palmer TJ, Rebolj M, Castañón A, Soldan K, Sasieni P. Trends in cervical cancer in women vaccinated against HPV: analysis of national cancer registry data. BMJ. 2024;386:g1001. doi:10.1136/bmj.g1001

---

Serious Harms and Adverse Events

5. Hviid A, Hansen JV, Frisch M, Mølgaard-Nielsen D. HPV vaccination and risk of autoimmune diseases: a population-based study. J Intern Med. 2021;289(4):480–489. doi:10.1111/joim.13156

6. Andersson K, Sundström K, Ploner A, et al. Risk of chronic fatigue syndrome after HPV vaccination: population based cohort study. BMJ. 2023;381:e076581. doi:10.1136/bmj-2022-076581

---

CIN2+/CIN3+ Outcomes Without Genotyping

7. Brotherton JML, Budd A, Rompotis C, et al. Is one dose of human papillomavirus vaccine as effective as three? BMJ Glob Health. 2019;4(2):e001245. doi:10.1136/bmjgh-2018-001245

8. Rebolj M, Mathews C, Castañón A, Sasieni P. HPV vaccination and cervical screening in England: lessons from analysis of 5.2 million cytology tests. Int J Cancer. 2022;151(4):575–584. doi:10.1002/ijc.34041

---

Type Replacement and Prevalence Studies

9. Tota JE, Ramanakumar AV, Jiang M, et al. Evaluation of type-specific HPV prevalence changes after vaccination in Canada. J Infect Dis. 2021;224(5):791–800. doi:10.1093/infdis/jiab271

10. Herweijer E, Sundström K, Ploner A, Uhnoo I, Sparén P. Quadrivalent HPV vaccine effectiveness against high-grade cervical lesions by age at vaccination: a population-based study. J Natl Cancer Inst. 2016;108(1):djv366. doi:10.1093/jnci/djv366

COI Source Documentation Links

1. ESGO–EFC Position Paper (2019)

2. ESGO–EFC Cervical Screening Consensus (2020)

3. Medscape CME Module co-authored by Markowitz

4. HPV Prevention and Control Board

5. Papillomavirus Research 2017 Board paper

6. BMC Proceedings Board report (2019)

7. NCT03979014 – NOVEL Trial (Kyrgiou, PI)

8. Cochrane Database Syst Rev 2018; CD009069

9. Cochrane Database Syst Rev 2025; CD015363

10. WHO SAGE HPV Vaccine Evidence Reports

11. EUROGIN Congress Programme

12. CDC MMWR HPV Vaccine Recommendations