Cat Fleas on Trial: Scientists Push Back on Claims That Fleas Spread Feline Hemoplasmas

A simmering dispute over whether the common cat flea can transmit hemotropic mycoplasmas—cell-wall-less bacteria that infect feline red blood cells—has erupted into print in the journal Parasites & Vectors. In a formal reply published as an open-access letter, a team of researchers led by Charlotte Moore and Edward Breitschwerdt of North Carolina State University, together with Erin Lashnits of the University of California Davis and Michael Lappin of Colorado State University, has defended its earlier meta-analysis against a published critique, while conceding and correcting specific data errors. The exchange matters far beyond academic point-scoring: hemoplasmas cause anemia in cats worldwide, and the question of how these bacteria move between animals shapes everything from veterinary screening advice to flea-control recommendations.

The controversy began when the group published a systematic review and meta-analysis in 2024 examining evidence that Ctenocephalides felis, the cat flea, carries or transmits hemotropic Mycoplasma species, including Mycoplasma haemofelis and the smaller Candidatus Mycoplasma haemominutum. Their analysis reached a provocative conclusion: when researchers used species-specific primers targeting hemoplasma genes, hemoplasma DNA turned up in fleas only rarely, and the widely used Jensen primers—developed in 2001 to detect what was then called Haemobartonella felis—may actually be amplifying DNA from Spiroplasma, a related bacterium that naturally inhabits fleas. A subsequent author responded with a critique alleging methodological and interpretive flaws, and the new letter is the team’s detailed rebuttal.

The reply opens with an unusually candid acknowledgment. The authors thank their critic for identifying mistakes in the data collection underlying the meta-analysis and have issued corrected figures in their supplementary tables for three studies—those by Zarea and colleagues, Abdullah and colleagues, and Shaw and colleagues. They also accept a specific miscategorization: the Shaw study did, in fact, wash fleas before DNA extraction, and it followed polymerase chain reaction testing with sequencing in 16 of 44 flea DNA extracts from cats, confirming a match to M. haemominutum over a 193-base-pair fragment of the 16S rRNA gene. An updated table now reflects that correction. Yet the team argues that none of these fixes overturns its central conclusion.

The reasoning is technical but consequential. Even though the Shaw study washed its fleas—something that should, in principle, remove surface contaminants—roughly 40 percent of the fleas, 44 of 90, still tested positive with the Jensen primers. To the authors, that pattern strengthens rather than weakens their hypothesis: if washing does not reduce positivity, the amplified signal may come from bacteria residing within the flea itself, such as Spiroplasma species, rather than from hemoplasmas acquired in a blood meal. The Jensen and Manvell primer sets, which were grouped in the meta-analysis, share a reverse primer and differ by only a single base pair in the forward primer—a shift that is homologous among feline hemoplasmas and the Spiroplasma species of the cat flea. In the original Manvell study, numerous initial PCR positives were reclassified after sequencing revealed them to be Spiroplasma rather than hemoplasma, a cautionary tale about what 16S rRNA primers with insufficient specificity can pick up.

For the studies that used non-Jensen, species-specific primers, the corrected numbers barely move the needle. In the Zarea study, the proportion is ambiguous because two Ctenocephalides canis fleas removed from cats fell outside the inclusion criteria, and the original authors did not specify whether all seven hemoplasma-positive fleas were cat fleas; a conservative estimate is 7 of 91, a prevalence under 8 percent. In the Abdullah study, applying the inclusion criteria strictly means counting only flea pools collected from cats—210 pools—and, within those, only the two positive pools with a confirmed feline host, rather than the originally reported 3 of 467. Recalculating with 2 of 210, about 1 percent, makes no statistically meaningful difference. Across all studies using species-specific primers, the updated overall proportion of hemoplasma DNA amplified from cat fleas stands at 14 of 949, or 1.5 percent—a figure the authors describe as evidence that such detection remains exceedingly rare.

A second flashpoint concerns pooling. Fleas are often processed in groups for DNA extraction, and pool size directly affects the probability of detecting low-abundance bacterial DNA, yet few studies report how many fleas each pool contained. The authors concede that inconsistent pool sizes complicate meta-analysis and urge future researchers to report pool composition comprehensively. In the Willi study, 73 fleas from 17 cats were processed in pools of one to five fleas, and although the results section stated that all positive samples came from individually extracted arthropods, the discussion suggested that pooling might have lowered the apparent prevalence—an internal tension that makes the study impossible to classify cleanly as pooled or unpooled. In the Abdullah study, at least half of the supposed pools were single fleas, with no further pool-size information available. Even so, the authors note, the bottom line survives: with species-specific primers, only two individual fleas in the Willi study and two pools—likely but not definitively single-flea pools—out of more than 200 in the Abdullah study were positive.

The experimental heart of the original paper involved retesting samples from a 2012 study by Assarasakorn and colleagues in Bangkok, which had reported a high prevalence of hemoplasma DNA in fleas using Jensen primers. The reanalysis found a dramatically lower prevalence, and the critic suggested sample degradation as a possible explanation. The authors counter that a substantial body of research shows DNA remains stable when stored frozen for years, and they address the pooling critique with new detail. According to Lappin, a co-author of the original Bangkok study, fleas there were pooled by cat—many cats contributing a single flea, with a maximum of five per pool—and when cats carried more than five fleas, multiple pools were created but only one was tested. That practice explains the numerical discrepancy between the 50 pools tested originally and the 67 pools from the same 50 cats retested in the new work. All 67 original DNA extracts were retested at Colorado State University with Jensen primers and at North Carolina State University with Manvell primers. Of the 17 previously untested pools, 7—41 percent—were positive with Jensen primers, compared with 18 of 50, or 36 percent, of the originally tested pools, results now laid out explicitly in a new table.

On the washing experiments, the authors acknowledge a genuine limitation: because the hemoplasma status of the fleas was unknown, the effect of washing on detection could not be determined—a caveat they admit could have been emphasized more forcefully. Fleas with known hemoplasma infection were simply not available, and those experiments remain to be done. This concession, however, is paired with a firm defense of the broader evidence base. The experimental infection studies cited in the original review involved small numbers of animals, but the authors argue those numbers are appropriate given how labor-intensive it is to work with cat fleas and cats in controlled settings, and at the time of publication the results represented the only published experimental transmission study using fleas.

Fresh evidence from outside the dispute has since tilted the picture further. A new experimental study published after the original manuscript found less support for flea transmission and more for direct transmission of hemoplasmas. In that work, hemoplasma transmission occurred between flea-infested cats sharing an enclosure—one of six cats infected with M. haemofelis and three of six with M. haemominutum—but no transmission was observed between cats kept in separate enclosures that allowed flea movement but no direct contact. By contrast, Bartonella clarridgeiae, a bacterium from a genus known for arthropod-borne spread, transmitted among all flea-infested cats in the same enclosure and to two of eight cats in adjacent enclosures permitting only flea movement, demonstrating a far clearer vector role for the cat flea. Geography adds another layer of skepticism toward a flea-driven model: in the Swiss survey, higher hemoplasma prevalence in cats occurred in the south and west of the country, yet the alpine south has a subarctic to tundra climate, not the warm, humid conditions that favor flea abundance. Surveys in Japan, spanning subarctic to humid subtropical climates, and Italy, from tundra to hot-summer Mediterranean zones, likewise fail to tie hemoplasma prevalence to flea-friendly conditions, and both included flea-borne Bartonella surveillance that allows direct comparison.

The authors close by clarifying their intent: they never sought to dismiss the high hemoplasma prevalences reported with Jensen primers, but to caution that those primers may not be as hemoplasma-specific as originally believed—a concern reinforced by recent reports of incidental amplification of Wolbachia and other flea microbiome DNA with inadequately specific 16S rRNA primers. They accept that fleas could yet prove to be competent vectors, but argue that geographic surveys and the newest experimental evidence more strongly support direct transmission among cats, through biting, grooming, or fighting. The path forward, all parties seem to agree, runs through better experiments: standardized washing protocols, transparent pool-size reporting, sequencing confirmation with rigorous BLAST analysis, and head-to-head comparisons of conventional PCR, quantitative PCR, and other molecular methods. Until then, the true route by which these elusive bacteria colonize feline blood remains an open and actively contested question.

Subject of Research: Vector competence of the cat flea for feline hemotropic Mycoplasma species and the specificity of molecular detection methods

Article Title: A case of misleading conclusion: a critical reassessment of methodological and interpretive flaws of a recent meta-analysis on the vector role of cat fleas in feline hemotropic Mycoplasma species

Article References: Moore, C. O., Lashnits, E., Lappin, M., & Breitschwerdt, E. B. (2026). A case of misleading conclusion: a critical reassessment of methodological and interpretive flaws of a recent meta-analysis on the vector role of cat fleas in feline hemotropic Mycoplasma species. Parasites & Vectors, 19(1), Article 372. https://doi.org/10.1186/s13071-026-07315-2

Image Credits: AI Generated

DOI: 10.1186/s13071-026-07315-2

Keywords: cat flea, hemotropic Mycoplasma, Ctenocephalides felis, meta-analysis, PCR primer specificity, Spiroplasma, Mycoplasma haemofelis, vector-borne disease, feline hemoplasmosis, Ctenocephalides canis, 16S rRNA, direct transmission