Antimalarial and antiplasmodial compounds (2024–2026): a systematic review of synthetic, semi-synthetic, and natural agents

Resistance to available antimalarial drugs continues to emerge and spread and remains a major challenge for global health. This systematic review summarizes the progress made in the discovery of antimalarial and antiplasmodial compounds published from January 2024 to June 2026, including synthetic, semi-synthetic, and hybrid molecules; natural products; and computational and delivery technologies. A systematic search for primary record identification was performed in Scopus, Web of Science, PubMed/MEDLINE, EMBASE, Google Scholar Alerts, ClinicalTrials.gov, and WHO sources according to PRISMA 2020 guidelines. Only supplementary full-text retrieval and citation tracing were conducted using publisher/journal platforms. The search strategy and available audit information are provided in Supplementary File S1. The number of historical retrievals and database-specific search dates that could not be independently verified were not estimated. We excluded preprints that were not later published in peer-reviewed journals afterward. Only primary research with original antimalarial/antiplasmodial data was included quantitatively, while reviews were used for background and citation chasing. One reviewer extracted data using a piloted standardized form, with independent verification of 20% of studies by two co-reviewers. Stratified by design, a credibility appraisal was applied. Out of 252 eligible primary studies, 157 (62.3%) were natural product-derived compounds, 49 (19.4%) synthetic compounds, and 46 (18.3%) semi-synthetic or hybrid molecules. Key mechanistic themes include inhibition of hemozoin (β-hematin) formation, dihydrofolate reductase, histone deacetylases, kinases, and proteasomes, as well as host-directed strategies. Mechanisms were separated based on experimental validation vs. computational prediction. 84/252 studies (33.3%) evaluated chloroquine-resistant parasites, 19/252 (7.5%) artemisinin/K13-associated resistance, and 9/252 (3.6%) ACT partner-drug resistance. Several compounds demonstrated nanomolar potency in vitro, but few had supporting in vivo, pharmacokinetic, or toxicological data. Antimalarial discovery is increasingly integrating medicinal chemistry, molecular hybridization, computational approaches, host-directed therapies, and nanoparticle-based delivery. However, most candidates are still at an early stage of development and are not yet proven solutions to drug resistance. Further in vivo validation, resistance testing, pharmacokinetic and toxicological characterization, and independent replication are necessary prior to clinical development.

Malaria, caused by protozoan parasites of the genus and transmitted by mosquitoes, remains a major global public health burden. According to the World Health Organization (WHO), malaria was estimated to have caused 282 million cases and 610,000 deaths worldwide in 2024. The WHO African Region continues to bear the greatest burden, accounting for approximately 94% of all malaria cases and 95% of malaria-related deaths globally. The emergence and spread of reduced susceptibility to artemisinin in several African countries poses a growing threat to malaria control efforts (1).

Pharmacological treatment of malaria has evolved through successive generations of antimalarial drugs, including cinchona alkaloids (quinine), 4-aminoquinolines (chloroquine), antifolates (sulfadoxine–pyrimethamine), and, since the 1990s, artemisinin-based combination therapies (ACTs) (2, 3). The efficacy of these first-line therapies is increasingly compromised by partial resistance to artemisinin, primarily associated with mutations in the Kelch13 (PfK13) gene, and by resistance to ACT partner drugs (4–6). This evolving epidemiological challenge has prompted intensive global efforts to discover and develop new antimalarial agents with novel mechanisms of action, improved pharmacokinetic properties, and efficacy against drug-resistant P. falciparum strains (7).

Contemporary antimalarial drug discovery advances along three complementary directions. First, synthetic medicinal chemistry exploits privileged heterocyclic scaffolds for the rational design of compounds with enhanced potency, selectivity, and target engagement (8–10). Second, semi-synthetic and hybrid approaches optimize established pharmacophores—particularly artemisinin and its derivatives—through molecular hybridization to improve efficacy, metabolic stability, and multi-target activity (11–13). Third, natural products remain a valuable source of structurally diverse lead compounds, with ongoing investigation of bioactive metabolites and plant extracts from ethnomedicinal species native to Africa, Asia, and South America, as well as marine organisms and endophytic microorganisms (14–16).

Computational approaches are now integral to each of these strategies. Molecular docking, quantitative structure–activity relationship (QSAR) modeling, molecular dynamics (MD) simulations, machine learning, artificial intelligence (AI), and network pharmacology are increasingly used to identify molecular targets, prioritize lead compounds, predict pharmacological properties, and guide structural optimization. It is important to note, however, that these in silico outputs constitute hypotheses about mechanism or target engagement rather than experimentally validated proof; this distinction is maintained throughout the present review (10, 17–20).

This systematic review provides an up-to-date synthesis of the antimalarial and antiplasmodial primary-research literature published between January 2024 and June 2026 (21–23). Two hundred and fifty-two eligible primary studies were included and analyzed according to chemical class, source organism or molecular scaffold, type of biological evaluation, resistance phenotype tested, therapeutic strategy, geographic origin, and translational maturity.

The specific objectives of this review were to: (i) systematically identify and synthesize primary studies reporting antimalarial and antiplasmodial compounds published between January 2024 and June 2026; (ii) classify these compounds into synthetic, semi-synthetic/hybrid, and natural product-derived categories, and characterize their chemical scaffolds, source organisms, biological activities, and mechanisms of action; (iii) critically evaluate reported potency, selectivity, and resistance-relevant activity against drug-sensitive and drug-resistant strains using explicit, predefined criteria; (iv) examine the contribution—and the epistemic limits—of computational approaches to antimalarial drug discovery; and (v) identify recurrent molecular targets, resistance-testing gaps, geographic patterns, and priority areas for future antimalarial drug discovery and translational development.

Review design

The study-selection process is summarized in the PRISMA 2020 flow diagram (Figure 1) and was conducted in accordance with the PRISMA 2020 Statement (24). This review aimed to systematically identify, evaluate, and synthesize primary studies published between January 2024 and June 2026 that reported antimalarial and antiplasmodial compounds of synthetic, semi-synthetic/hybrid, or natural-product origin, with particular attention to chemical scaffolds, mechanisms of action, nanoparticle-based drug-delivery systems, computational drug-discovery approaches, and activity against drug-resistant strains.

The research question is phrased using the PICOS framework adapted for a preclinical drug-discovery review: Population/Problem— in vitro, in vivo or in silico models of (and where reported, other species) relevant to human malaria; Intervention – antimalarial/antiplasmodial compounds of synthetic, semi-synthetic/hybrid or natural-product origin, including nanoparticle-based formulations; Comparison—where reported, reference antimalarial drugs (e.g. chloroquine, artemisinin/dihydroartemisinin, atovaquone), drug-sensitive parasite strains or untreated/vehicle controls; Outcome: Antiplasmodial/antimalarial activity (e.g., IC 50 /EC 50, parasitemia reduction, or survival); selectivity and cytotoxicity; mechanism of action; and activity against resistance phenotypes. Study design: Original research articles reporting in vitro, in vivo, in silico, or translational research (including formulation and drug-delivery studies), as defined in Section 2.3.

The review was not prospectively registered in PROSPERO or any other systematic-review registry. This is recognized as a methodological limitation because prospective registration can reduce the risk of post hoc changes to the review question or eligibility criteria, and this limitation is discussed further in Section 4.

Literature search strategy

Zotero audit clarification: the supplied Zotero report was used to verify bibliographic metadata for records retained in the review and to identify item-level publication and consultation/access dates. It does not preserve the historical query strings or database-level retrieval counts, so those fields are not presented as if they had been recovered.

Seven sources were used for primary record identification: Scopus, Web of Science, PubMed/MEDLINE, EMBASE, Google Scholar Alerts, ClinicalTrials.gov, and WHO websites/reports. Publisher/journal platforms (ScienceDirect, SpringerLink, Wiley Online Library, MDPI, ACS Publications, Nature Portfolio, Taylor & Francis Online, and Frontiers) were used as supplementary full-text/citation-tracing sources and were not counted as independent record-identification sources. The strategy combined Medical Subject Headings (MeSH), where supported, with free-text terms related to malaria, antimalarial compounds, medicinal chemistry, natural products, semi-synthetic derivatives, molecular hybrids, nanoparticle-based drug-delivery systems, computational drug discovery, drug resistance, and Plasmodium falciparum, using the Boolean operators “AND” and “OR.” The search concepts and available search-audit information are provided in Supplementary File S1. The original database-level search-history exports were not retained for all sources. The Zotero report available for this resubmission preserves item-level bibliographic metadata, publication dates, and consultation/access dates for individual records, but it does not preserve the original database query strings, database-level hit counts, or a complete database-level search-date history. These historical fields are therefore reported as NR (not retained/not recoverable), rather than reconstructed or estimated. Supplementary File S1 distinguishes the historical search-audit fields that are verifiable from the bibliographic information available in the Zotero report. Reference lists of eligible articles and relevant reviews were additionally screened manually (citation chasing) to identify potentially eligible primary studies. Reviews identified in this way were not themselves entered into the 252-study corpus (see Section 2.3). The historical number of studies contributed exclusively by citation chasing was not retained in a separately exportable field and is therefore reported as NR in Supplementary File S1 rather than reconstructed from the final corpus.

Eligibility criteria

Primary research studies were eligible for inclusion in the quantitative corpus if they met all of the following criteria:

published between January 2024 and June 2026 (inclusive);

written in English;

original research articles (in vitro, in vivo, in silico, or translational) reporting new experimental or computational data on synthetic, semi-synthetic, natural-product-derived, molecular-hybrid, or nanoparticle-formulated antimalarial/antiplasmodial compounds;

reported at least one quantitative or clearly described qualitative measure of antiplasmodial/antimalarial activity, mechanism of action, or resistance phenotype.

Studies were excluded if they were:

duplicate publications or duplicate/overlapping datasets;

conference abstracts without an accessible full text;

editorials, letters, commentaries, or opinion pieces;

not related to antimalarial or antiplasmodial drug discovery;

Preprints were not eligible as independent records in the primary-study corpus. Preprints without subsequent peer-reviewed publication were excluded. When a preprint was subsequently published as a peer-reviewed article within the eligibility period, the peer-reviewed publication was retained as the eligible record and the preprint was not counted separately.

Narrative and systematic reviews on antimalarial drug discovery published in the same window were identified during the search (n = 34) but were not included in the 252-study primary corpus; they were used only for background context and citation chasing and are cited as such (e.g., refs (9, 21, 22, 25, 26, 50, 51, 101, 102, 108).) rather than counted as original evidence. This separation was introduced specifically to avoid double-counting of evidence between reviews and the primary studies they cite (see Section 4, limitations).

Because in vitro, in vivo, in silico, and translational studies differ fundamentally in the strength of inference they support, these study types are labelled explicitly throughout the Results and are compared, not pooled, in the synthesis (Section 2.6).

Study selection

Records retrieved from the database searches were imported into a reference-management system, and duplicates were removed. Titles and abstracts were screened against the eligibility criteria above. Full-text articles considered potentially eligible were retrieved and independently assessed for inclusion by two reviewers (G.F.N.N. and H.T.N.N.); disagreements were resolved by discussion with the third reviewer (B.M.). The study-selection process is summarized in the PRISMA 2020 flow diagram (Figure 1).

Data extraction

Data were extracted using a standardized extraction form piloted on a random sample of 15 studies before full-scale use. Full extraction of all 252 included studies was performed by a single reviewer (B.M.), given resource constraints; to mitigate single-reviewer extraction bias, a random 20% sample (n = 50; rounded down from 50.4 studies) was independently re-extracted by the second and third reviewers (G.F.N.N., H.T.N.N.), and discrepancies were resolved by consensus. Extracted fields included: year of publication; compound or molecular scaffold; chemical class (synthetic, semi-synthetic/hybrid, or natural-product-derived); biological source; species and strain(s), including resistance genotype/phenotype where reported; experimental approach (in vitro, in vivo, in silico, translational); mechanism of action and whether it was experimentally validated or only computationally predicted; reported potency (IC50/EC50) and, where available, selectivity index or cytotoxicity data; nanoparticle-based formulation details; computational analyses performed; country/region of the source organism and of the corresponding author’s institution; and the principal findings. This single-reviewer extraction with partial independent verification is acknowledged as a methodological limitation (Section 4).

Operational definitions and data synthesis

Owing to substantial heterogeneity in study designs, compound classes, experimental models, and outcome measures, a quantitative meta-analysis was not feasible; findings were synthesized narratively. To reduce interpretive drift, the following operational definitions were applied consistently when describing activity:

“Highly potent”: in vitro IC50/EC50 ≤ 100 nM against at least one P. falciparum strain.

“Potent”: IC50/EC50 in the range >100 nM–1 μM.

“Moderately active”: IC50/EC50 in the range >1–10 μM.

“Selective”: selectivity index (mammalian cytotoxicity IC50/antiplasmodial IC50) ≥ 10, only applied when cytotoxicity data were reported in the primary study.

“Promising” is used only where a study reported both potency and either a resistant-strain comparator or in vivo /selectivity data; where only bare in vitro potency against a single sensitive strain was reported, this is stated explicitly rather than described as “promising”

“Preclinical relevance/maturity” is graded using the three-tier translational framework in Section 3.9 (hit, lead, preclinical candidate), not inferred from potency alone.

The included studies were classified into three principal categories—synthetic compounds, semi-synthetic/hybrid molecules, and natural product-derived compounds—and, within and across these categories, evidence was comparatively analyzed with respect to chemical scaffolds, molecular targets, mechanisms of action (validated vs. predicted), antiplasmodial activity (using the thresholds above), resistance profiles, nanoparticle-based delivery systems, computational approaches, and combination strategies, in order to identify emerging trends, knowledge gaps, and priorities for future antimalarial drug discovery.

Quality assessment and risk-of-bias methodology

The corpus included multiple fundamentally different types of study design (in vitro pharmacology, in vivo rodent efficacy, in silico/computational prediction, and translational/formulation studies), and there was no single validated risk-of-bias instrument (e.g., Cochrane RoB2, ROBINS-I, or QUADAS-2, all developed for clinical intervention or diagnostic-accuracy studies) that was directly applicable to the entire dataset. Instead of the narrative-synthesis approach outlined in Section 2.6, we employed a design-stratified quality-assessment method, adapting elements from four established frameworks to the study type at hand. This appraisal is an explicit qualitative/semi-quantitative aid to interpretation and not a formally validated risk of bias tool. This distinction is made explicit to the reader.

In vitro pharmacology studies (synthetic, semi-synthetic, and natural-product screening): assessed with a checklist adapted from CASP (Critical Appraisal Skills Program) principles for laboratory pharmacology, scoring (a) inclusion of a positive reference/control compound, (b) reporting of assay replication (≥ 3 biological or technical replicates, or explicit statement of n), (c) reporting of a selectivity or cytotoxicity measure against a mammalian cell line, and (d) sufficient methodological detail (assay type, exposure time, strain identity) to permit replication.

In vivo studies (rodent models): evaluated against the relevant items of the ARRIVE 2.0 guidelines, namely, randomization to treatment group; blinding of parasitemia assessment; sample-size justification or a priori sample-size statement; and reporting of animal housing/ethical-approval information.

The evaluation criteria for computational studies were adapted to the specific methodological approach used. The assessment of QSAR studies was performed in accordance with the OECD principles for the validation of (Q)SAR models, which comprise the definition of the endpoint, applicability domain, robustness, predictivity, and mechanistic interpretation. Domain-specific criteria were applied to molecular-docking and molecular-dynamics studies, including the use of appropriate controls, validation of docking protocols, assessment of binding-site and ligand interactions, reproducibility of computational procedures, and experimental corroboration of predicted targets or interactions, where available. These criteria were used to assess the methodological quality and credibility of computational findings, rather than to provide a formal risk-of-bias score.”

Translational/nanoparticle-formulation studies: evaluated for (a) physicochemical characterization of the formulation (size, polydispersity, encapsulation efficiency or equivalent); (b) stability data beyond a single time point; and (c) an appropriate unformulated/free drug comparator.

Natural product studies specifically (crossing over the above categories where applicable): additionally scored for whether the source plant/organism was formally authenticated and a voucher specimen deposited in a named herbarium or culture collection, as a proxy for taxonomic reliability of the reported source.

Each study was assessed for overall quality as low, moderate, or high, based on the number of criteria applicable to its design category that the study met (high = all/nearly all criteria met, moderate = roughly half, low = few or none). The primary reviewer performed the appraisal, and an independent reviewer double-checked a random 20% sample using the same verification procedure as described for data extraction in section 2.5. Available study-level quality-audit records are shown in Supplementary Table S2. These records represent 74 of the 252 included studies and are therefore a partial audit trail. They are not a full reconstruction of the appraisal at the level of the study. Where individual criterion-level records were not retained in the archived material, they are reported as NR rather than reconstructed.

Data analysis

Publication trends, compound classes, therapeutic targets, resistance-testing coverage, geographic origin, and translational maturity were summarized using descriptive analyses. Findings are presented as a narrative synthesis supported by summary tables and figures.

Overview of the corpus

A total of 252 primary publications published between January 2024 and June 2026 met the eligibility criteria and were included in the qualitative synthesis (Figure 1). Of these, 157 studies (62.3%) investigated natural product-derived compounds, 49 studies (19.4%) focused on synthetic compounds, and 46 studies (18.3%) evaluated semi-synthetic or hybrid molecules (Table 1; Figure 2).

The initially reported publication-year counts did not add up to 252 (5 + 109 + 66 + 25 = 205, short by 47 studies); this was due to “online-first” articles with month-only dates being counted as a separate additional category rather than being included within their final publication year and to a residual set of accepted-but-not-yet-issue-assigned papers being excluded from the year-by-year total entirely. This has been corrected, and the reconciled breakdown is shown in Table 2 below, which by construction sums to exactly 252.

Table 2 presents the counts for the entire 252-study corpus at the year level according to the final PRISMA screening and extraction records. Supplementary File S1 is a separate audit subset derived from Zotero and was not intended to reproduce the publication-year distribution of the full corpus. The additional export of the Zotero library was used only as a bibliographic cross-check and did not substitute the original screening and eligibility records.

A later raw Zotero library export was used only as an external bibliographic cross-check and was not treated as the original PRISMA screening database. Because the later export does not preserve the original full-text eligibility decisions, it was not used to modify the final 252-study corpus or its publication-year classification. The final corpus and temporal distribution therefore remain based on the documented PRISMA screening and extraction records described in this review.

Publication activity therefore increased markedly from 2024 to 2025 and remained high through the first half of 2026; because 2026 data cover only six months, this partial-year count cannot be directly compared with the complete years 2024–2025, and no extrapolated annual estimate is presented.

The journals contributing the largest number of publications were J Ethnopharmacol (n = 7), J Med Chem (n = 6), Nat Prod Res (n = 4), Eur J Med Chem (n = 3), Pharmaceuticals (Basel) (n = 3), Malar J (n = 3), PLoS One (n = 3), Sci Rep (n = 3), and Pharmacol Res – Nat Prod (n = 3).

For design-stratified quality appraisal (Section 2.7), 61 studies (24.2%) were rated as overall high quality, 148 (58.7%) moderate, and 43 (17.1%) low; the available study-level audit records are provided in Supplementary Table S2, while the aggregate distribution is reported here. The breakdown by study design category and individual criteria (e.g., ARRIVE items for in vivo work, external validation of computational studies, and voucher-specimen deposition for natural-product studies) is reported in full in the new Section 3.8. This distribution is used as a reference to interpret the strength of particular claims throughout the Results section.

Synthetic compounds

The 49 studies classified as synthetic compounds reported a diverse range of heterocyclic and non-heterocyclic molecular scaffolds, with quinoline derivatives and their hybrids the most extensively investigated chemical class, consistent with recent reviews highlighting quinoline-based hybrids as privileged scaffolds for antimalarial optimization (25, 26) (background reviews, not part of the 252-study corpus).

Recent primary studies include aminopropyl-linked quinoline-triazole hybrids with enhanced antiplasmodial activity attributed to inhibition of β-hematin (hemozoin) formation and novel bisquinoline derivatives inspired by classical 4-aminoquinoline scaffolds for drug-resistant malaria (27, 28). Quinazoline derivatives also emerged as candidates, including substituted 2-(4-(sulfonyl)piperazin-1-yl)quinazoline hybrids and 4-benzyloxy-2-trichloromethylquinazoline derivatives designed as dihydrofolate reductase (PfDHFR) inhibitors on the basis of docking/QSAR modeling; target engagement in these two studies was computationally predicted and has not yet been confirmed by direct biochemical (e.g., enzyme-inhibition) assays (29, 30).

Histone deacetylase (HDAC) inhibitors have been investigated as epigenetic modulators of parasite development, including pentyloxyamide-based HDAC inhibitors with experimentally confirmed antiplasmodial activity, alongside improved QSAR models for hydroxamic-acid HDAC inhibitors (31, 32). Naphthoquinone derivatives and metal–atovaquone complexes were also explored for antiplasmodial activity and structure–activity relationships (33, 34). Other reported synthetic hits included 1,2,3-triazoles derived from propargyl alcohol (35), spiroindoline phenotypic hits (37), thymol derivatives with ADMET profiling (36), 2-pyrazoline carboxamide derivatives (39), arylpyrrole derivatives (38), and fluorinated cannabidiol/cannabigerol derivatives (40); potency for most of these compounds falls in the moderately active to potent range (>100 nM–10 μM) as defined in Section 2.6, with only a minority (see below) reaching the highly potent (≤100 nM) threshold, and cytotoxicity/selectivity data were reported for fewer than half of these studies. 5,6,7,8-Tetrahydroimidazo[1,2-a]pyridine (THIP) derivatives were also reported, including compounds with single-digit-to-low double-digit nanomolar potency against the asexual blood stage (IC50 = 8.7 nM), which meets our highly potent threshold; the study also reported a favorable early selectivity index against a mammalian cell line, but no in vivo efficacy or pharmacokinetic data were available, so preclinical relevance remains unconfirmed (41).

Innovative delivery and target-based strategies included polyethylene glycol–lumefantrine conjugates to improve solubility and pharmacokinetics and dual PfA-M1/PfA-M17 aminopeptidase inhibitors with cross-species activity (42, 43). Computational approaches—in silico gold(I)-based hybrids targeting hemozoin formation, QSAR modeling, and phenotypic screening—were extensively integrated into lead identification and optimization, although, as above, computational target predictions in these studies await biochemical confirmation (32, 37, 44). Green nanotechnology approaches, including plant-mediated MgO and ZrO 2 nanocomposites, expanded the diversity of nanomaterial-based strategies (45).

Semi-synthetic and hybrid compounds

The 46 studies classified as semi-synthetic and hybrid compounds emphasized artemisinin-based derivatives, molecular hybridization, and nanotechnology-enabled delivery. A major focus was the structural optimization of artemisinin and dihydroartemisinin (DHA): dendrimer-based dual delivery co-encapsulating artesunate and fluconazole showed enhanced efficacy in a murine malaria model (in vivo, moderate-to-high credibility rating) (46), and oligochitosan conjugates incorporating DHA and lumefantrine were developed as multifunctional delivery systems (47).

Additional chemical diversity included alkoxy-substituted spiro[indoline-3,1′-isoquinolin]-2-one derivatives (48) and steroid–tetraoxane hybrids characterized by consensus molecular docking and molecular dynamics simulations—again, a computationally predicted, not yet biochemically validated, interaction profile (49). Artemisinin-based hybrid molecules and second-generation β-carboline/chloroquine hybrids showed experimentally confirmed antiplasmodial activity with reported structure–activity relationships (53, 54). Curcumin derivatives were evaluated against chloroquine-resistant P. falciparum using combined experimental/computational approaches (55), and nanoparticle-based curcumin delivery systems (e.g., chitosan–hyaluronic-acid-functionalized polycaprolactone nanoparticles) were developed to improve bioavailability (57). Alstonine and related β-carboline scaffolds are discussed as background in the literature (56), but no pre-2024 study is used here to represent evidence from the 2024–2026 primary corpus.

Another cluster was the plant-mediated green synthesis of metal/metal-oxide nanoparticles: MgO nanoparticles synthesized with Achillea millefolium extract were active against chloroquine-resistant P. falciparum (58); silver nanoparticles from Piper umbellatum extract showed antiplasmodial activity (59); ZnO nanoparticles from Eucalyptus camaldulensis showed antioxidant and antiplasmodial activity (60). Other hybridization strategies involved α,β-unsaturated ketones derived from thiazole developed as potential antimalarial agents, for which combined biophysical and computational data support the inhibition of Plasmodium falciparum lactate dehydrogenase (PfLDH) as a likely mechanism of action. Since this conclusion is based on docking and biophysical (not purified enzyme activity) data, PfLDH inhibition should be regarded as a computationally supported hypothesis, rather than a validated enzymatic finding, pending a direct enzyme-inhibition assay (61).

SAR patterns combined with docking also led to assignment of mechanisms for quinoline-ferrocine-tetrahydropyrimidine-triazole hybrids (62) and heterocyclic chalcone derivatives (63), and these target assignments also await direct biochemical confirmation. Reverse fosmidomycin analogs have been included as bisubstrate inhibitors in structure-based semi-synthetic work (64), semi-synthetic lupane triterpenoid (betulin, betulinic acid, and lupeol) derivatives with QSAR-guided optimization (65, 66), and genome-guided mining of cyclolipodepsipeptides (67).

Natural product-derived agents

Natural product-derived agents formed the largest category (n = 157), reflecting both the biodiversity available for antiplasmodial phytochemistry and the continued relevance of ethnopharmacological knowledge in malaria-endemic regions. Plant extracts, purified phytochemicals, marine natural products, and microbial metabolites were evaluated using in vitro, in vivo, in silico, or combined methodologies.

Artemisia remained the most frequently studied genus. Studies compared the antiplasmodial activity and phytochemical composition of A. annua teas using untargeted metabolomics, examined the influence of extraction method on antiplasmodial potency, assessed A. afra and A. annua aqueous extracts against artemisinin-resistant P. falciparum, and investigated the pharmacodynamic/pharmacokinetic basis of the A. annua whole-plant matrix effect (68–72). Cadinapyridine sesquiterpene alkaloids from A. annua were characterized for cytotoxic and antiplasmodial activity, and agronomic interventions (intercropping, microbial soil amendments) were shown to potentiate antimalarial activity in cultivated A. annua (73, 74).

Curcuma longa was also extensively studied: curcumin poly(D,L-lactide) nanoparticles showed superior antiplasmodial efficacy relative to DHA in an integrated in vivo /in silico analysis, hydroethanolic leaf extract showed in vitro activity, and curcumin displayed multiple proposed effects including PfGSK3 inhibition and cytokine modulation (75, 76). Vernonia species showed activity in rodent models (V. amygdalina against P. berghei) and in vitro against P. falciparum and Trypanosoma species; combined bark extracts of V. amygdalina, Mangifera indica, and Carica papaya (1:1:1) were active against chloroquine-resistant strains (77–79). Alstonia species (A. scholaris, A. boonei) showed antiparasitic and immunomodulatory activity (80, 81). Compounds from Solanum nigrum, including Dobinin K, showed activity against chloroquine-resistant P. falciparum via proposed membrane-disruption mechanisms, identified through a combination of in vitro assay and network-pharmacology prediction (82, 83).

Numerous African medicinal plants yielded additional leads, including 28-norlupanes and dammaranes from Dipterocarpus costatus (IC50 = 3.74 μM against chloroquine-resistant P. falciparum, within our moderately active range) (88), secondary metabolites from Pleiocarpa bicarpellata (84), antiplasmodial/antioxidant constituents from Haematostaphis barteri (85), neo-clerodane furanoditerpenoids from Croton macrostachyus (87), bioactive fractions of Landolphia owariensis with β-hematin inhibition (86), and compounds from Phragmanthera capitata and Tetracera alnifolia (89, 90). Peptide and alkaloid natural products included the peptaibols trichokonins VI and VIII (91), a genome-guided overview of antimalarial cyclolipodepsipeptides (67), and the bisindole alkaloid cyclosebistine A from Catharanthus roseus (92).

In silico and computational natural-product studies included docking-based identification of candidate plasmepsin II inhibitors from neem phytochemicals (93), network-pharmacology assessment of CE-326597 (94), and adaptive docking characterisation of putative FabI (enoyl-ACP reductase) inhibitors from Salacia debilis (95); in each case, the reported target interactions are computational predictions that have not yet been confirmed by direct biochemical assay. Experimentally supported findings included quercetin inhibition of P. falciparum small heat-shock proteins (96) and the repurposing of the oncology ATM-kinase inhibitor AZD0156 as a P. falciparum PI4Kβ inhibitor, which was confirmed biochemically (97). Nanoparticle formulations of natural products included microencapsulated Carica papaya leaf extract, quinine-loaded chitosan–polycaprolactone nanoparticles, and silver nanoparticles synthesized using Padina tetrastromatica and Piper umbellatum extracts (47, 98–100).

Mechanistic targets and modes of action

A common set of mechanistic targets recurred across all three categories (Figure 3). Inhibition of β-hematin (hemozoin) formation was the most frequently reported mechanism, particularly for quinoline derivatives, naphthoquinones, and plant alkaloids, and was experimentally supported (e.g., β-hematin inhibition assays) in the majority of studies reporting it. Pentyloxyamide scaffolds and hydroxamic acids targeted HDAC inhibition with biochemical confirmation. Quinazoline derivatives were designed against PfDHFR primarily on computational grounds, as noted in Section 3.2. Proteasome inhibition has been discussed as an emerging antimalarial strategy in recent reviews (101, 102), but these reviews were not counted among the 252 primary studies and are not used here as primary evidence. Kinase inhibition — including PI4Kβ and Rac1 in eligible recent primary studies — was identified as an important strategy with biochemical or genetic validation in the corresponding primary studies (103). Earlier evidence supporting PKG as an antimalarial target is discussed separately as historical background (104). Glutathione S-transferase and single-stranded DNA-binding protein were investigated as candidate targets, the latter via a novel scaffold series active against both drug-sensitive and artemisinin-resistant strains (105). Inhibition of the fosmidomycin pathway (MEP isoprenoid biosynthesis) was pursued as a validated antimalarial strategy (64). Atovaquone–metal complexes and tafenoquine were reported to target the parasite mitochondrion via electron-transport interference. Host-directed approaches, including cholesterol-sulfate modulation and Rac1 inhibition, represent an emerging but still mechanistically less-validated category, discussed further in Section 3.7 (103, 106).

Computationally informed mechanisms were assessed at the study level rather than by counting individual mechanistic claims. Among the 96 studies containing a computational component, 39 (40.6%) reported independent biochemical or biophysical corroboration of the proposed target or interaction, whereas computational outputs without such corroboration were treated as hypothesis-generating rather than as validated mechanism or target engagement. These proportions are defined at the study level: among the 96 computational studies, 39 (40.6%) had independent biochemical or biophysical corroboration of the proposed target or interaction and 57 (59.4%) did not. They therefore do not represent the proportion of individual mechanistic claims, compounds, targets, or mechanisms.

Alongside this validated/predicted split, the computational studies in this corpus also varied greatly in their internal methodological rigor. Of the 96 studies (38.1% of the 252-study corpus) that reported a docking, QSAR, molecular-dynamics, or network-pharmacology analysis, only 34 (35.4%) described some form of external validation (independent test set, leave-one-out cross-validation, or similar) using the OECD criteria outlined in Section 2.7. An applicability-domain statement was reported in 21 studies (21.9%), and a negative, scrambled-ligand, or randomization control was reported in only 14 studies (14.6%). It is also known from the wider computational chemistry literature that docking binding-energy scores correlate imperfectly with true binding affinity when viewed in isolation and are sensitive to protein-structure preparation and choice of scoring function (17–20). Therefore, favorable docking scores in this corpus should be viewed as hypothesis-generating rather than confirmatory in the absence of independent validation. We emphasize this point explicitly here because computational and experimental methods, although often referred to as complementary or “inseparable” in the source literature, differ considerably in the degree of causal inference they afford, and conflating them risks overstating the maturity of the underlying evidence.

Drug resistance considerations

Based on the extraction database, 84 of 252 studies (33.3%) explicitly tested compounds against chloroquine-resistant (CQ-resistant) P. falciparum strains or, in in vivo studies, resistant parasite lines; this was the most common resistance model used for synthetic small molecules, including quinoline and curcumin derivatives (55, 82, 107), and for several nanotechnology-based interventions (ZnO and green-synthesized MgO nanoparticles and plant extracts) (33, 59). Nineteen studies (7.5%) tested activity specifically against artemisinin-resistant or PfK13-mutant parasite lines (70, 105), and nine studies (3.6%) examined activity relevant to ACT partner-drug resistance. The remaining studies either did not test against a defined resistant phenotype or tested only against laboratory-sensitive reference strains (e.g., 3D7, NF54). This distribution indicates that, while resistance is a stated motivation for most of the reviewed work, direct experimental testing against artemisinin/K13-associated resistance—the resistance phenotype of greatest current clinical concern (1, 4–6)—remains comparatively limited within this corpus.

Host-directed strategies represent an alternative approach to circumventing conventional resistance mechanisms: cholesterol-sulfate modulation via repurposing of the steroid sulfatase inhibitor irosustat was shown to disrupt intraerythrocytic development of P. falciparum through a host-directed mechanism distinct from currently available parasite-targeted drug classes (106). Earlier experimental work on parasite PKG inhibitors demonstrated prophylactic, blood-stage, and transmission-blocking activity (104). This pre-2024 study is cited solely as historical mechanistic background and is not used as evidence from the 2024–2026 primary corpus. Recent narrative reviews of resistance mechanisms and mitigation strategies are cited for broader context but, consistent with Section 2.3, are not counted among the 252 primary studies (101, 102, 108, 109).

Complementary and synergistic compound combinations

A recurring and increasingly intentional theme in the 2024–2026 primary literature is the design, formulation, or empirical testing of compound combinations with complementary or synergistic antiplasmodial activity. To avoid conflating distinct pharmacological concepts, these are separated into four categories below. Note that “complementary mechanism” (two agents acting on different molecular targets) is not the same claim as demonstrated pharmacological synergy (a supra-additive combined effect, typically shown by isobologram analysis or a combination index); unless a study explicitly reports such a synergy metric, we describe the relationship as “mechanistically complementary” rather than “synergistic”.

Molecular hybrids (single covalent scaffold, two pharmacophores)

β-Carboline–artemisinin conjugates (“harmisinins”) are among the most extensively discussed hybrid scaffolds in the literature reviewed here (52–54), addressed across three publications describing harmisinins as antiplasmodial/cytotoxic agents, β-carboline–artemisinin hybrid synthesis, and second-generation harmiquins. The proposed rationale—β-carboline-mediated DNA intercalation/hemozoin inhibition combined with artemisinin endoperoxide-mediated radical alkylation—is mechanistically plausible and supported by SAR data in the primary studies, but supra-additive synergy (as opposed to additive dual activity) was not formally quantified (e.g., via combination index) in any of the three publications; because one of these studies predates the review window, it is cited as background only; we therefore describe this as an orthogonal dual-mechanism hybrid rather than a demonstrated synergistic combination. Quinoline–triazole and quinoline–ferrocene–tetrahydropyrimidine–triazole hybrids constitute a second major class, with proposed complementary hemozoin-inhibition and redox-cycling/iron-chelation activities (28, 62, 110); again, this complementarity is mechanistic rather than a quantified synergy claim. Steroid–tetraoxane hybrids similarly combine a membrane-active triterpenoid scaffold with an artemiside-class endoperoxide, characterized so far only by consensus docking/MD (computational prediction) (49).

Co-formulations and covalent drug conjugates

A dendrimer-based dual-delivery system co-encapsulating artesunate and fluconazole showed in vivo combination efficacy in a murine model (46); because the two agents are physically co-delivered rather than covalently hybridized, and because fluconazole’s contribution is proposed to act via host immunomodulation/pharmacokinetic effects rather than a directly overlapping antiplasmodial mechanism, this is best described as pharmacokinetic/pharmacodynamic complementarity supported by in vivo (not merely in vitro) evidence—among the stronger examples in this corpus. Oligochitosan conjugates of DHA and lumefantrine (47) and polyethylene glycol–lumefantrine conjugates (42) were designed to reproduce the differential-release, stage-sequential rationale of existing DHA–lumefantrine ACT formulations; both remain at the formulation/ in vitro stability stage without reported comparative in vivo efficacy data at the time of publication.

Polyherbal combinations

The only explicitly tested multi-extract polyherbal combination identified in this corpus was reported by Ebong et al.: combined bark extracts of Vernonia amygdalina, Mangifera indica, and Carica papaya (1:1:1) tested against chloroquine-resistant P. falciparum in a murine model (78). The proposed mechanistic basis – hemozoin-inhibiting sesquiterpene lactones from Vernonia amygdalina, antioxidant/antiprotozoal mangiferin from M. indica, and membrane-targeting isothiocyanates/alkaloids from C. papaya – is plausible but it has not been independently confirmed for each constituent within the combined-extract study, thus it is reported here as a single empirically tested polyherbal formulation rather than a combination for which the individual mechanisms of its constituent extracts have been experimentally disentangled.

Host-directed and parasite-directed dual targeting

A conceptually distinct form of complementary strategy arises from pairing host-directed and parasite-directed agents. Carbazole–pyrimidines targeting the host protein Rac1 (103), the repurposed kinase inhibitor AZD0156 (PI4Kβ) (97) each act, at least in part, on pathways distinct from conventional parasite metabolic targets. Earlier PKG inhibitor work (104) is cited separately as historical background only and was not included in the 2024–2026 primary corpus. Because host-directed agents do not share a target with conventional ACT-partner drugs, resistance to the latter would not be expected to confer cross-resistance to the former; however, this expectation has not yet been experimentally tested against clinically resistant isolates for any of the agents identified in this corpus and should be regarded as a testable hypothesis rather than an established finding. Irosustat, a repurposed steroid sulfatase inhibitor that modulates host cholesterol-sulfate levels, follows the same host-directed logic (106). The main complementary target pairs identified are summarized in Table 3, with the strength-of-evidence caveats above applying to each row.

Design-stratified quality and credibility appraisal

Aggregate design-stratified quality and credibility results are reported for the 252-study corpus using the methodology described in Section 2.7.

The study designs were heterogeneous (in vitro, in vivo, in silico/computational, translational/formulation, and natural-product studies), and thus design-specific criteria were used instead of a single generic risk-of-bias instrument. Results are summarized below by study design category. Quality-audit information is provided in Supplementary Table S2 as an aggregate audit summary. The complete historical study-by-study criterion matrix for all 252 included references was not retained, so S2 should not be interpreted as a full study-level reconstruction. The aggregate percentages reported below were calculated from the full review dataset at the point of data extraction and quality appraisal and cannot be independently recalculated from S2 alone. Where criterion-level records were not retained, values are reported as NR rather than reconstructed. In vivo studies (n=41 studies containing a rodent-model component): randomization to treatment group was explicitly reported in 19 studies (46.3%); blinded parasitology assessment was reported or implied in 11 studies (26.8%); a priori sample-size justification was reported in 6 studies (14.6%); and a statement of ethical/institutional animal-care approval was present in 33 studies (80.5%). Thus, compliance with the full ARRIVE 2.0 checklist was partial and uneven, with randomization and blinding—arguably the items most relevant to safeguarding against bias—least consistently reported.

In silico/computational studies (n = 96 studies containing a docking, QSAR, molecular-dynamics, or network-pharmacology component): an external validation was reported in 34 (35.4%), an applicability-domain statement in 21 (21.9%), and a negative/scrambled control in 14 (14.6%) of the 96 studies, as detailed in Section 3.5. The proposed target was corroborated by an independent biochemical or biophysical assay in 39 of the 96 studies (40.6%), indicating that independent corroboration was present in 40.6% of the computational subset; this should not be interpreted as a direct estimate of all mechanistic claims.

Translational/nanoparticle-formulation studies (n=38 studies): Physicochemical characterization of the formulation was reported in 35 (92.1%); stability data beyond a single time point were reported in 22 (57.9%); and an appropriate unformulated/free-drug comparator was included in 24 (63.2%).

Natural-product studies specifically (n = 157): formal authentication of the source plant/organism with deposition of a voucher specimen in a named herbarium or culture collection was reported in 71 (45.2%); the remaining 86 studies (54.8%) did not report authentication or referred only to informal or non-deposited identification, which we consider an important and likely under-appreciated source of uncertainty in this literature, as misidentification of source material would invalidate downstream structure-activity conclusions.

The application of these design-specific criteria to the aggregate low/moderate/high categories reported in Section 3.1 (61 high, 148 moderate, 43 low) shows that the most common reason for a “low” rating was, in order of descending frequency: lack of a reference/control compound or comparator (in vitro / in vivo), lack of any external validation (in silico), and lack of botanical/source authentication (natural products). The results are described qualitatively throughout Sections 3.2–3.7 to temper claims of “promising” or “potent” activity when the underlying study was rated low on the relevant criteria, and are presented quantitatively in Table 4.

Translational maturity of candidate agents and interventions

To provide a clearer translational framework, each compound, formulation, or intervention discussed above was classified, based on the data extracted, into one of three tiers: (i) Hit— in vitro activity only, against a single (usually sensitive) strain, without selectivity or in vivo data; (ii) Lead— in vitro activity against at least one resistant strain and/or reported selectivity index, but without in vivo efficacy or pharmacokinetic data; (iii) Preclinical candidate— In vivo efficacy shown in an animal model; enough data to support further pre-clinical development. Table 5 lists representative examples from each tier; consistent with the overall corpus composition, the large majority of agents and interventions identified in this review (an estimated 70–75%, based on the extraction database) remain at the Hit or Lead stage, and only a small minority have reported in vivo efficacy data, underscoring that this literature represents an early-stage discovery pipeline rather than a set of clinically translatable candidates.

Geographic distribution of natural-product research

Given the review’s emphasis on African, Asian, and Amazonian biodiversity, the geographic origin of source organisms and the institutional affiliation of corresponding authors were extracted for the 157 natural-product studies. Source organisms were of African origin in 61 studies (38.9%), Asian origin in 54 studies (34.4%), South/Central American origin in 19 studies (12.1%), and marine/multiple/unspecified origin in 23 studies (14.6%). Corresponding-author institutions were located in Africa for 58 studies (36.9%), Asia for 71 studies (45.2%), Europe/North America for 21 studies (13.4%), and elsewhere for 7 studies (4.5%). Notably, a majority of studies on African-sourced plants were nonetheless corresponding-authored by African institutions (39 of 61, 63.9%), indicating substantial local research leadership in this domain, though international collaboration and capacity constraints in translational (in vivo /pharmacokinetic) work remain evident, since only 8 of the 61 African-sourced-plant studies (13.1%) reported in vivo efficacy data.

The corpus of 252 primary studies reviewed here reveals several defining themes in contemporary antimalarial drug discovery. First, the predominance of natural-product studies (62.3%) reflects the ongoing search for biodiversity-derived leads in malaria-endemic countries and the increasing, though still uneven, integration of ethnopharmacological knowledge into evidence-based pipelines (Section 3.10). The abundance of crude-extract and early-stage phytochemical studies, while valuable for biodiversity documentation, highlights the continuing challenge of converting natural-product hits into fully characterized drug candidates through isolation, structural elucidation, and mechanistic validation; in our translational tiering (Section 3.9), the great majority of natural-product studies remain at the hit or lead stage.

Second, the quinoline scaffold and its hybrids with triazole, ferrocene, chalcone, and aminoguanidine units remain the most productive privileged structure in synthetic antimalarial chemistry, reflecting both its historical link to chloroquine and its versatility for rational modification. The nanomolar potency of THIP derivatives (IC50 = 8.7 nM (41)) and of aminopropyl-linked quinoline–triazole hybrids (28) shows that even classical frameworks can still yield candidates with strong in vitro potency; however, as noted in Section 3.2, in vivo efficacy and pharmacokinetic data for these specific compounds were not available at the time of the reviewed publications, so their preclinical relevance cannot yet be confirmed on potency data alone.

Third, the semi-synthetic/hybrid class illustrates the continued centrality of the artemisinin pharmacophore. β-Carboline–artemisinin, steroid–tetraoxane, and spiro[indoline-isoquinolinone]-related hybrids represent a coherent strategy to combine artemisinin’s radical-mediated mechanism with a second, mechanistically distinct pharmacophore; as discussed in Section 3.7, this mechanistic complementarity is well supported by SAR data but has rarely been confirmed as quantitative pharmacological synergy. The genome-guided discovery of antimalarial cyclolipodepsipeptides (67) illustrates a promising convergence of biosynthetic-pathway elucidation and semi-synthetic library generation.

Fourth, computational drug discovery is now integral to experimental programs across all three classes; dozens of studies have applied QSAR modeling, molecular docking, MD simulations, network pharmacology, and machine learning (e.g., the TransQSAR-Pf framework integrating P. falciparum stress-response signatures (111)). We emphasize, however, that computational outputs are best regarded as prioritization tools and testable hypotheses rather than as evidence of validated mechanism or target engagement unless independently corroborated; among the 96 computational studies, 39 (40.6%) reported independent biochemical or biophysical corroboration (Section 3.5).

Fifth, growing interest in host-directed therapies, drug repositioning, and nanotechnology-based delivery reflects recognition that a parasite-only, single-target paradigm is likely insufficient and that pharmacokinetic innovation is as important as pharmacodynamic potency. Studies on PI4Kβ inhibition, irosustat repurposing as a host-targeted antiplasmodial, and nanoparticle-based delivery systems exemplify this multi-axis approach (97, 106); as discussed in Section 3.7.4, the theoretical resistance-robustness of host-directed strategies remains a hypothesis awaiting direct testing against clinically resistant isolates.

Sixth, the high prevalence of natural-product studies with positive antiplasmodial activity (Section 3.4) should be interpreted cautiously because publication bias and other forms of non-reporting can make studies with less favorable findings harder to identify in a systematic review (112). In this context, the positive-study profile observed here may overstate the underlying hit rate of ethnopharmacological screening. In this corpus, the risk is compounded by the finding that fewer than half of natural-product studies (45.2%; Section 3.8) reported formal authentication of the source plant material, introducing a further source of uncertainty, largely unquantifiable. We therefore caution against interpreting the 62.3% natural-product studies in this review as evidence that natural sources are disproportionately more likely than synthetic or semi-synthetic approaches to yield active antimalarial compounds; it more plausibly reflects a combination of genuine biodiversity-derived opportunity, low barriers to screening crude extracts, and asymmetric publication of positive results.

Seventh, on drug resistance specifically (Section 3.6), only a minority of studies in this corpus directly tested compounds against artemisinin/K13-associated or ACT-partner-drug resistance phenotypes, even though this is the resistance profile of greatest current clinical concern (1, 4–6). This is an important gap: potency against chloroquine-resistant parasites, while a useful and historically standard benchmark, does not by itself establish activity against the resistance mechanisms currently spreading in the WHO African Region.

This review has several limitations. The search strategy was restricted to English-language publications that were accessible to the review team, which may introduce publication and language bias; grey literature (conference proceedings, dissertations, and unpublished studies) was not systematically searched, which may limit comprehensiveness. Given the heterogeneity in study design, species/strains, experimental methods, and outcome measures, a formal meta-analysis could not be performed, and a single validated risk-of-bias tool could not be applied across all study types; instead, we used a design-stratified credibility appraisal (Section 2.7), which is a structured qualitative aid rather than a formal bias-quantification instrument. Full-text data extraction for all 252 studies was performed predominantly by a single reviewer, with independent verification limited to a 20% random sample; this remains a source of potential extraction error or selective emphasis despite the piloting and verification steps taken, and readers should weigh this when relying on specific extracted values. The review was not prospectively registered. Finally, because narrative and systematic reviews identified during the search were deliberately excluded from the 252 studies’ primary corpus to avoid double-counting (Section 2.3), this review may under-represent evidence that exists only in secondary/review form; we consider this an acceptable trade-off for the sake of avoiding inflated or duplicated evidence counts Plasmodium.

Future systematic reviews in this area should include a wider range of bibliographic databases and multilingual publishing strategies, systematic grey literature searching, prospective registration (e.g., PROSPERO), dual independent full-text data extraction for the entire corpus, and formal application of design-appropriate risk-of-bias tools (e.g., SYRCLE for animal studies) to increase methodological rigor and reduce potential bias.

This systematic review of 252 primary publications on antimalarial and antiplasmodial compounds from 2024 to 2026 provides a structured and, following revision, more critically appraised snapshot of the current state of malaria drug discovery. The corpus is quantitatively dominated by natural-product research, spanning African, Asian, and Amazonian medicinal plants as well as marine organisms and microbial metabolites—a substantial biodiversity reservoir for future drug leads, with encouraging evidence of research leadership from African institutions on African-sourced organisms (Section 3.10), though translational (in vivo /pharmacokinetic) capacity in this space remains comparatively limited. Synthetic medicinal chemistry based on quinoline, quinazoline, HDAC-inhibitor, triazole, and spiroindoline scaffolds has delivered candidates with nanomolar in vitro potency, and semi-synthetic/hybrid approaches continue to exploit the artemisinin pharmacophore and triterpenoid frameworks in novel combinations.

Mechanistically, hemozoin inhibition, HDAC and kinase inhibition, DHFR targeting, proteasome inhibition, and host-directed strategies together constitute a multitarget pharmacological landscape that reflects the complexity of biology. Computational tools—QSAR, docking, molecular dynamics, machine learning, and network pharmacology—are increasingly integrated across all three chemical classes and are accelerating candidate prioritization, but, as emphasized throughout this revised manuscript, computational predictions of mechanism or target engagement should be treated as hypotheses pending experimental confirmation, not as established findings Plasmodium.

Importantly, the evidence assembled here demonstrates potential rather than proof: no study in this corpus reports clinical-stage or clinically validated evidence that these approaches have overcome antimalarial drug resistance, and, based on our translational tiering (Section 3.9), the vast majority of identified compounds remain at the hit or lead stage. Key priorities for future research include the following: (i) systematic isolation and structural characterization of active principles from the many crude-extract studies identified; (ii) rigorous evaluation of the most promising synthetic and hybrid candidates in validated artemisinin/K13-resistant parasite models, not only chloroquine-resistant models, and in well-designed in vivo efficacy studies; (iii) experimental validation of computationally predicted mechanisms and targets; (iv) translation of nanoparticle and conjugate delivery platforms toward preclinical toxicology and bioavailability studies; and (v) continued and strengthened integration of African and tropical biodiversity research—including local institutional leadership—into global antimalarial drug-discovery pipelines. The 2024–2026 literature surveyed here reflects a field of considerable vitality and innovation, meriting continued translational investment, tempered by a clear-eyed accounting of how much work remains between an active compound in a dish and a viable antimalarial medicine.