Reviewers : Diptarup Mallick, Mwafaq Ramzi Haji, & Hala Taha Elbashir. Synthesized by Diptarup Mallick
Brief summary:
This manuscript investigates the effect of mixed Plasmodium vivax (Pv) and Plasmodium falciparum (Pf) infections on parasite density, gametocyte carriage, and transmission to mosquitoes in Ethiopia. The study enrolled 511 malaria patients across four health facilities and used qPCR/RT-qPCR for parasite and gametocyte quantification together with Direct Membrane Feeding Assays (DMFAs) to assess infectivity to Anopheles arabiensis. The authors report that Pf gametocyte prevalence and density were significantly lower in mixed infections, whereas Pv gametocyte prevalence remained high regardless of infection type. Importantly, mosquito infectivity was not significantly affected by mixed-species infection status after adjusting for gametocyte density. The study concludes that mixed infections contribute substantially to malaria transmission and are likely underestimated by routine diagnostics. The manuscript addresses an important and understudied aspect of malaria epidemiology and provides valuable transmission data from a co-endemic setting. The integration of molecular diagnostics with mosquito infectivity assays is a major strength.
This cross‑sectional study used qPCR and direct membrane feeding assays to compare asexual parasite densities, gametocyte densities, and mosquito infectivity between Plasmodium vivax mono‑infection, P. falciparum mono‑infection, and mixed‑species infections in Ethiopian patients, finding lower P. falciparum gametocyte density in mixed infections, uniformly high P. vivax gametocyte prevalence, and no evidence that co‑infection alters the relationship between gametocyte density and mosquito infection rates after adjustment for gametocytemia.
Major comments:
1. Clarification of mixed infection definition and diagnostic thresholds
The manuscript states that PCR was used for final infection classification because microscopy missed many mixed infections. However, the Ct thresholds, positivity cutoffs, and criteria for defining mixed infections are not clearly described. Since low-density infections are central to the conclusions, the manuscript would benefit from:
Explicit qPCR positivity thresholds
Clarification on how borderline or low-copy mixed infections were handled
Whether contamination controls were assessed quantitatively
This is especially important because mixed infections may include minority species at very low densities.
2. Cross-sectional design limits mechanistic interpretation
The authors appropriately acknowledge that the study is cross-sectional and cannot determine parasite kinetics or causal interactions. However, some sections of the Discussion still imply biological competition or suppression mechanisms between Pf and Pv. For example, statements regarding reticulocyte competition and anemia-mediated suppression are somewhat speculative.
The manuscript would be strengthened if:
Speculative mechanisms were more clearly labeled as hypotheses
Alternative explanations (host immunity, treatment-seeking behavior, infection timing, or seasonality) were discussed more explicitly
3. Statistical methods require additional detail
The statistical analysis section is concise but lacks important details needed for reproducibility. Specifically:
Were corrections for multiple comparisons applied?
How were missing data handled?
Were site-level effects evaluated?
Was mosquito batch or feeding experiment included as a random effect?
Because data were collected from multiple sies over several years, temporal and geographic heterogeneity may influence transmission patterns.
4. Interpretation of mosquito infectivity results
The conclusion that there is “no evidence for competition within mosquitoes” is interesting and important. However, this conclusion should be interpreted cautiously because:
Mosquito infection was assessed mainly through oocyst prevalence and PCR detection
The study did not directly measure sporogonic development success, salivary gland invasion efficiency, or transmission fitness beyond infection prevalence
Mosquito immune responses were not assessed
The authors should moderate the conclusion slightly or clarify that they found no detectable reduction in infectivity under the conditions tested.
5. Data availability statement is incomplete
The manuscript states that data and code “will be made available” and references a future GitHub link. The authors should provide:
An active repository link before publication
Accession information or DOI if available
More precise information about what datasets and scripts will be shared
Cross‑sectional design limits causal inference on interspecies interactions.
The authors interpret lower gametocyte densities in mixed infections as potential evidence of blood‑stage competition (e.g., via reticulocyte limitation or dyserythropoiesis). However, all data derive from a single time point. Without longitudinal follow‑up of the same infections, it is impossible to exclude that the observed differences simply reflect the natural history (e.g., different ages of infection, asynchronous peaks) or unmeasured confounders (e.g., prior antimalarial use, host immunity). The Discussion acknowledges this limitation, but the text repeatedly infers “suppression” or competitive effects (e.g., “this finding appears consistent with prior observations that co‑infection may suppress P. falciparum gametocyte carriage”). I recommend toning down causal language throughout and explicitly stating that cross‑sectional associations cannot establish directionality.
Unbalanced sample sizes and potential selection bias.
The analysis hinges on comparisons between 284 Pv mono‑infections, 150 Pf mono‑infections, and only 77 mixed infections (of which 48 were used in feeding assays). The heavy imbalance raises concerns about statistical power, particularly for interaction analyses (e.g., Spearman correlations in mixed infections reported as non‑significant for Pf, p=0.570, and borderline for Pv, p=0.120). Were these null results truly due to absent association or simply a consequence of small sample size in the mixed‑infection group? Confidence intervals for the correlation coefficients should be presented. Moreover, participants were enrolled by microscopy, which misclassified 48% of mixed infections as mono‑infections (Table 1). How might this initial selection have biased the composition of the mono‑infection groups? The true “mono” groups may contain cryptic mixed infections, which would dilute differences. A sensitivity analysis restricting to PCR‑confirmed mono‑infections alone would strengthen the findings.
Missing details on confounder adjustment and model building.
The GAMM analysis (Fig 3, S1 Table) is central to the claim of no competition within mosquitoes. The models included log₁₀ gametocyte density and infection type (mono/mixed) as fixed effects, with a random intercept for individual. Were other potential confounders — site, season, age, fever status, asexual parasitemia — tested? These could influence both gametocyte density and mosquito infectivity independently of co‑infection status. Including them, or at least demonstrating that their omission does not alter the key result, would greatly strengthen the conclusion. Additionally, the model for P. falciparum used total gametocyte density (sum of male and female), but gametocyte sex ratio is a known determinant of infectivity; sex‑specific densities should be examined.
Interpretation of the correlation between asexual and gametocyte densities is problematic.
The Abstract states: *“Statistically significant positive correlations … were observed in mono‑infections (p<0.001), but not in mixed‑species infections (p=0.120 for Pv; and p=0.570 for Pf).”* This is taken as evidence that co‑infection decouples gametocyte production. However, a non‑significant p‑value does not demonstrate equivalence or absence of correlation. Moreover, Fig 3A and 3B appear to show scatterplots that might still show a positive trend; formal interaction tests between gametocyte density and infection type should be performed (e.g., via linear models with an interaction term) rather than separate correlations. Without this, the claim of a biological “decoupling” is unjustified.
The “hidden risk” framing is not fully substantiated by the data.
The final sentence of the Abstract asserts: “Because mixed infections are often undetected, they represent a hidden risk for sustaining malaria transmission.” While it is true that microscopy misses many mixed infections, the study did not compare transmission potential between detected and undetected mixed infections, nor did it quantify the population‑level contribution of undetected mixed infections relative to mono‑infections. This statement should be tempered, or the necessary epidemiological modelling should be provided.
Incomplete reporting of experimental procedures and data.
Several methodological details essential for reproducibility are omitted or unclear:
How were the standard curves for P. vivax qPCR validated? Using plasmid constructs assumes equal amplification efficiency, which should be verified against cultured parasites of known concentration.
The DMFA used colony‑maintained An. arabiensis, but wild mosquito populations may differ in vector competence. The authors previously demonstrated comparability for P. vivax (ref 20), but mixed infections were not specifically tested. A comment on this limitation is warranted.
Sporozoite data (S2 Fig) are described as “comparable”, but no statistical test is reported. Provide the appropriate test and exact p‑values.
The numbers of mosquitoes dissected for oocysts varied (81 mosquitoes total for some feeds?). Reporting variability in mosquito numbers per feed and how it was handled in the binomial GAMM is important.
The Discussion should more fully explore potential blood‑stage interaction mechanisms. The data show that P. vivax gametocyte density was lower in mixed infections despite no apparent vector‑stage competition. The authors should expand the mechanistic discussion—for example, elaborating on how P. falciparum‑induced anaemia and dyserythropoiesis may limit reticulocyte availability for P. vivax—to better contextualise these patterns and avoid the impression that the absence of mosquito‑stage competition negates the possibility of blood‑stage competition.
Minor comments:
1. Typographical and grammatical issues
Several typographical errors should be corrected:
“ther production” → “their production”
“expecially” → “especially”
“falciaprum” → “falciparum”
“gametcoyte” → “gametocyte”
“Variabls” → “Variables”
A thorough language edit would improve readability.
2. Figure legends need improvement
Some figure legends are overly long and difficult to follow. Consider simplifying and separating methodological details from interpretive information.
3. Consistency in reporting parasite density units
Parasite densities are reported in different units (copies/µL, parasites/µL, transcripts/µL). While scientifically valid, more clear standardization or explanation would improve readability.
4. Reference formatting inconsistencies
Several references contain formatting inconsistencies or typographical errors. For example:
“Plasmodium vevax” instead of “vivax”
Some references lack punctuation consistency
A careful reference check is recommended.
Strengths of the Study:
Large sample size across multiple endemic regions
Combination of molecular diagnostics and DMFA transmission assays
Important contribution to understanding mixed-species malaria transmission
Strong public health relevance for malaria elimination programs
Well-structured overall narrative with biologically meaningful findings
Comments on reporting:
Correct a numerical inconsistency in patient characteristics. The text reports that “50.9% (220/511) were febrile,” but 220/511 = 43.05%. The authors should verify whether the numerator (febrile participants) or denominator (total enrolled) is misstated, or correct the percentage accordingly.
The preprint aligns with the STROBE checklist on core content (rationale, methods, participant characteristics, results, and limitations), but falls short on several structural items: sample size justification, participant flow diagram, explicit treatment of bias, detailed statistical methods including confounder adjustments and missing data, and a more rigorous generalisability discussion. At minimum, the authors should add a participant flow chart, a statement on bias, and clarify the statistical modelling approach.
Suggestions for future studies :
Longitudinal cohort design. A prospective study that follows the same individuals with repeated measures of asexual parasitaemia, gametocytemia, and mosquito infectivity over time is essential to distinguish competitive suppression from differences in infection age, to quantify gametocyte conversion rates, and to determine whether mixed infections alter infection longevity.
Experimental co‑infection models. Controlled human infection studies or ex‑vivo gametocyte culture systems infected with both P. falciparum and P. vivax could isolate the direct effects of one species on the other’s gametocyte production and maturation, removing confounders such as prior immunity or differential exposure.
Wild vector assessments. Repeating the study using freshly colonised or wild‑caught Anopheles from co‑endemic settings would strengthen the external validity, as laboratory‑bred mosquitoes may differ in vector competence and feeding behaviour.
Immunological and haematological profiling. Future work should measure markers of anaemia, reticulocyte count, and species‑specific immune responses (e.g., antibody levels, cytokine profiles) to mechanistically test the hypothesis that P. falciparum‑induced dyserythropoiesis restricts P. vivax parasitaemia in mixed infections.
Inclusion of asymptomatic and community‑based infections. Extending recruitment beyond symptomatic health‑facility attenders would capture the contribution of low‑density, asymptomatic mixed infections to the transmission reservoir—an important gap that the present study highlights but does not fill.
Multi‑site studies across diverse transmission settings. Conducting similar investigations in other co‑endemic regions with different vector species, drug‑resistance profiles, and transmission intensities would assess the generalisability of the finding that gametocyte density, rather than co‑infection per se, determines mosquito infection success.
Sex‑ratio and stage‑specific gametocyte analyses. Because the probability of mosquito infection depends on male‑female gametocyte ratios and maturity, future studies should examine whether co‑infection skews gametocyte sex ratios or arrests maturation, potentially explaining the loss of correlation between asexual and gametocyte densities in mixed infections.
Population‑level transmission modelling. Once longitudinal and community‑based data on mixed‑infections become available, mathematical models could help quantify the hidden transmission risk that undetected mixed infections pose to malaria elimination efforts, clarifying whether targeted molecular screening would be cost‑effective.
The authors declare that they have no competing interests.
The authors declare that they used generative AI to come up with new ideas for their review.