Researchers at a single center analyzed patients who received first-line gemcitabine plus either cisplatin or carboplatin between 2014 and 2024. The primary endpoint was the change in renal function during treatment, defined as baseline eGFR minus the lowest eGFR recorded during chemotherapy; a larger value therefore represented a greater absolute decline. Because renal measurements had skewed distributions, the investigators used non-parametric tests for comparisons between groups. Variables associated with the outcome in univariable analyses at p < 0.10 were entered into a multivariable linear regression.

The two treatment groups differed at baseline. Patients receiving gemcitabine plus carboplatin were older and had significantly lower baseline eGFR than those receiving gemcitabine plus cisplatin. They also had a smaller distribution of absolute eGFR decline in the univariable comparison. The univariable regression coefficient for regimen, expressed as gemcitabine plus cisplatin versus gemcitabine plus carboplatin, was −16.116, with a 95% confidence interval of −23.317 to −8.915 and p < 0.001. However, regimen was not independently associated with renal decline after baseline eGFR and the other eligible variables were included in the multivariable model.

Baseline eGFR had a strong positive correlation with the amount of renal-function decline across the cohort, with r = 0.794 and p < 0.001. The multivariable coefficient for baseline eGFR was 0.523, with a 95% confidence interval of 0.396 to 0.650 and p < 0.001, indicating that each unit of baseline eGFR was associated with an estimated 0.523-unit increase in the absolute decline. The investigators attributed the smaller unadjusted decline observed in the carboplatin group to treatment selection based on pre-existing renal impairment rather than to demonstrated differences in regimen-specific nephrotoxicity. The study did not establish that carboplatin was less nephrotoxic than cisplatin. Its retrospective design, single-center setting, unequal group sizes and small carboplatin subgroup also limit generalization. The results identify baseline renal function as the principal measured predictor of eGFR decline in this population, rather than providing evidence that either chemotherapy regimen prevents kidney injury or establishing which regimen an individual patient should receive.

Subject of Research: Baseline eGFR as a predictor of renal decline during platinum-based chemotherapy for advanced urothelial carcinoma

Article Title: Predictors of renal function decline during platinum-based chemotherapy in advanced urothelial carcinoma

Article References: Fukushima, K., Miyasaka, M., Shikano, R., Higuchi, S., Asada, S., & Hashimoto, T. (2026). Predictors of renal function decline during platinum-based chemotherapy in advanced urothelial carcinoma. BMC Cancer. https://doi.org/10.1186/s12885-026-16897-0

Image Credits: AI Generated

DOI: 10.1186/s12885-026-16897-0

Keywords: urothelial carcinoma, renal function, estimated GFR, cisplatin, carboplatin, platinum chemotherapy, nephrotoxicity, retrospective study, Predictors, renal, function, decline

The researchers collected demographic, clinical and laboratory data from inpatients treated with vancomycin between April 1, 2024, and February 28, 2025. Data were analyzed using SPSS, with statistical significance set at P < 0.05. The study was prospective and observational, so its reported associations should not be interpreted as proof that vancomycin or any accompanying factor directly produced the observed changes in kidney function.

According to the abstract, patients who had never attended school had an adjusted odds ratio (AOR) of 9.86 for acute nephrotoxicity compared with the reference group (P = 0.012; 95% CI, 1.67–58.16). Those who could read and write had an AOR of 11.49 (P = 0.013; 95% CI, 1.69–78.07). Urinary tract infection was associated with an AOR of 15.5 (P = 0.02; 95% CI, 1.49–161.02), while concomitant omeprazole use was associated with an AOR of 6.43 (P = 0.002; 95% CI, 2.01–20.54).

The abstract also reported lower odds of acute nephrotoxicity among housewives (AOR 0.11; P = 0.004; 95% CI, 0.03–0.50), farmers (AOR 0.04; P < 0.001; 95% CI, 0.007–0.18) and people working in private jobs (AOR 0.061; P = 0.004; 95% CI, 0.01–0.35) than the relevant reference group. These estimates come from a single-center cohort and should be read as reported associations rather than causal effects. The available source is an abstract published in BMC Infectious Diseases, with the article noting that the version is subject to further edits; the abstract does not provide the full methodological details, adjustment variables, subgroup results or limitations needed for a complete assessment of the findings.

The authors concluded that the findings highlight the importance of clinical awareness surrounding potentially high-risk medication use. They further state that renal-function screening should be a central component of care for vulnerable patients receiving vancomycin. The summary does not report an intervention trial, so it provides no direct evidence that a screening strategy reduces nephrotoxicity or improves clinical outcomes.

Subject of Research: Predictors of acute nephrotoxicity among vancomycin-treated inpatients in a prospective cohort study

Article Title: Assessment of nephrotoxicity risk and its predictors among vancomycin-treated patients admitted at Debre Markos comprehensive specialized hospital

Article References: Biazen, M. G., Mengist, A., Abebaw, A., Tegegne, B. A., Muche, Y., Biyazin, Y., Endalew, B., Meharie, B. G., & Elmneh, H. T. (2026). Assessment of nephrotoxicity risk and its predictors among vancomycin-treated patients admitted at Debre Markos comprehensive specialized hospital. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14491-0

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14491-0

Keywords: vancomycin, acute nephrotoxicity, serum creatinine, acute kidney injury, prospective cohort, inpatients, risk factors, omeprazole, urinary tract infection, Assessment, nephrotoxicity, risk

High-dose methotrexate is used in treatment regimens for acute lymphoblastic leukemia, lymphoma and other malignancies, but it has a narrow therapeutic window and can produce severe adverse effects. The drug is excreted unchanged through the kidneys, while a small portion is metabolized in the liver to 7-hydroxymethotrexate, which is also cleared renally but more slowly. In acidic environments, 7-hydroxymethotrexate has only one-third to one-fifth the solubility of methotrexate, making precipitation in renal tubules more likely. Such a process can cause renal injury, which in turn reduces methotrexate clearance and prolongs exposure to toxic concentrations. Oxidative stress and inflammatory responses are also associated with methotrexate-related liver and kidney damage, providing a rationale for examining the redox forms of homocysteine rather than focusing only on their combined concentration.

The researchers enrolled patients with acute lymphoblastic leukemia or lymphoma who received high-dose methotrexate between June 2024 and August 2025. Pediatric patients received doses of 2–5 grams per square metre under Chinese guidance, while adult regimens ranged from 1–5 grams per square metre, with 5 grams per square metre applicable to T-cell acute lymphoblastic leukemia. Methotrexate concentrations were monitored at 42 hours, and leucovorin rescue doses were adjusted accordingly. Rescue was given every six hours until methotrexate fell below specified concentrations, while hydration and urine alkalinization were used to support elimination. All participants had normal hepatic and renal indicators before treatment.

Blood was collected at 42, 72 and 96 hours after methotrexate administration, producing 310 samples from 89 patients: 38 adults and 51 children. Some patients contributed repeated samples within one treatment cycle or across different cycles. There were 143 samples at 42 hours, 110 at 72 hours and 57 at or beyond 96 hours. The number of renal-toxicity events at those times was 12, 17 and 11, respectively. Following the National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0, nephrotoxicity was defined as creatinine above the laboratory upper limit of normal. Hepatotoxicity required alanine aminotransferase above that limit together with either aspartate aminotransferase or total bilirubin above the limit. Laboratory abnormalities meeting these definitions and corresponding to at least grade 1 were classified as events.

Total homocysteine was measured in plasma with an enzymatic cycling method. The investigators separately quantified free reduced homocysteine and homocystine using ultrafiltration and HPLC-MS/MS. N-ethylmaleimide was added during sample preparation to prevent thiol auto-oxidation, an important analytical step because free reduced homocysteine is susceptible to conversion into its oxidized form. The assays were validated for calibration range, linearity, lower limits of quantification, precision, accuracy, recovery, matrix effects and stability. The lower limits were 0.05 micromoles per litre for the reduced-form assay and 0.004 micromoles per litre for homocystine; intra-day and inter-day precision had relative standard deviations below 10%, and matrix-effect coefficients of variation were also below 10%. The Hcy/HHcy concentration ratio was calculated as an indirect indicator of the relative redox state of free homocysteine.

Generalized estimating equations were used to account for repeated observations and examine associations between homocysteine measures and indicators of liver or kidney function. Models were adjusted for sex, age, height, body weight, baseline renal and hepatic indicators, methotrexate concentration, methotrexate dose and delayed elimination. False-discovery-rate correction was applied to multiple comparisons, and pediatric and adult patients were analyzed separately. Among children, Hcy/HHcy was significantly negatively associated with creatinine, with a regression coefficient of −0.007, a 95% confidence interval of −0.012 to −0.002 and an adjusted P value of 0.028. Total homocysteine was also negatively associated with alanine aminotransferase and total bilirubin in this age group. Its association with direct bilirubin had an FDR-adjusted P value of 0.065 and therefore did not meet the study’s corrected threshold for statistical significance. No significant associations between the homocysteine measures and hepatic or renal indicators were detected in adults.

Compared with patients whose kidney function remained within the defined laboratory range, those classified as having kidney injury had significantly different concentrations of reduced homocysteine and homocystine at 42 hours, with FDR-adjusted P values of 0.039 and 0.003, respectively. Homocystine was elevated in the kidney-injury group at that time. Reduced homocysteine differed between the groups at 72 hours, with an adjusted P value of 0.003, and homocystine differed at or beyond 96 hours, with an adjusted P value below 0.001. No significant differences in total homocysteine, reduced homocysteine, homocystine or the Hcy/HHcy ratio were found between the liver-injury and normal-function groups after adjustment. The clearest observed group differences therefore involved renal injury, but the observational comparisons do not establish that changes in homocysteine redox forms caused the kidney damage.

Exploratory receiver operating characteristic analyses examined how well the markers distinguished organ injury. Homocystine at 42 hours showed the strongest reported discrimination for renal injury, with an AUC of 0.817, a 95% confidence interval of 0.702–0.932 and a P value below 0.001. At the selected threshold of 0.04 micromoles per litre, sensitivity was 66.7% and specificity was 87.0%. Total homocysteine at 42 hours had an AUC of 0.681, with a 95% confidence interval of 0.492–0.871. The original AUC for Hcy/HHcy was 0.232, indicating that lower ratios were associated with renal injury; reversing the direction of that index produced an adjusted AUC of 0.768, with a 95% confidence interval of 0.628–0.908 and a P value of 0.002. Total homocysteine continued to discriminate renal injury at 72 hours and at or beyond 96 hours, with AUCs of 0.761 and 0.812, respectively.

By contrast, total homocysteine showed reverse discrimination for hepatotoxicity at 72 hours and at or beyond 96 hours, with AUCs of 0.654 and 0.715, meaning that lower concentrations were associated with liver injury in those analyses. Across all sampled observations, median total homocysteine was 8.53 micromoles per litre at 42 hours, 10.13 at 72 hours and 9.61 at or beyond 96 hours. Median homocystine was 0.017 micromoles per litre at the first two time points and 0.020 at or beyond 96 hours, while the median Hcy/HHcy ratio fell from 36.38 at 72 hours to 19.77 at the later point. These were cross-sectional summaries rather than continuous trajectories measured in every patient, because sample numbers decreased and patient overlap was incomplete across time points.

The authors proposed that methotrexate inhibition of dihydrofolate reductase could interfere with the folate-dependent remethylation of homocysteine to methionine, increasing homocysteine early in treatment. Recovery of the methionine cycle and calcium folinate rescue might then contribute to a later decline in total homocysteine, while continued oxidation of reduced homocysteine could increase homocystine and lower the Hcy/HHcy ratio. This mechanism remains hypothetical because oxidative stress was not measured directly. Important limitations further restrict interpretation: baseline homocysteine measurements were unavailable, the cohort was relatively small and clinically heterogeneous, and some potential confounders were not systematically collected. There was no internal validation of the ROC analyses, and the grade 1 toxicity threshold may have included mild or transient laboratory abnormalities. The study therefore provides preliminary evidence about homocysteine redox forms as potential indicators of methotrexate-associated nephrotoxicity, rather than establishing them as validated clinical biomarkers.

Subject of Research: Association of homocysteine redox forms with renal toxicity after high-dose methotrexate in leukemia and lymphoma

Article Title: The relationship of homocysteine and its redox forms with renal toxicity after high-dose methotrexate treatment in patients with leukemia or lymphoma

Article References: Gao, S.-S., Wu, C.-L., Liu, Z.-Y., Zheng, J.-X., Zhou, H.-Y., Xie, J.-Y., Guo, J.-L., & Dong, W.-C. (2026). The relationship of homocysteine and its redox forms with renal toxicity after high-dose methotrexate treatment in patients with leukemia or lymphoma. Frontiers in Pharmacology, 17, Article 1972047. https://doi.org/10.3389/fphar.2026.1972047

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1972047

Keywords: high-dose methotrexate, nephrotoxicity, homocysteine, homocystine, oxidative stress, redox ratio, leukemia, lymphoma, renal function, biomarkers, relationship, redox

The research, published in the journal Pharmacia, examined 44 patients with transfusion-dependent beta-thalassemia major and compared them with 30 healthy individuals matched for age and sex. Rather than relying on conventional serum markers alone, the investigators focused on two sensitive urinary indicators of early glomerular injury: the albumin-to-creatinine ratio, known as ACR, and the protein-to-creatinine ratio, or PCR. Both were measured in first-morning urine samples, a standard approach that minimizes day-to-day variation and captures the earliest signs of protein leaking through damaged kidney filters.

The results were statistically significant and clinically striking. Patients with beta-thalassemia major showed markedly higher ACR values than controls, averaging 6.94 compared with 0.63, a difference that reached significance at p = 0.015. PCR values followed the same pattern, averaging 20.46 in patients versus 9.38 in controls, significant at p = 0.023. In practical terms, this means that albumin and total protein were escaping into the urine of thalassemia patients at levels several-fold above normal, a hallmark of glomerular barrier dysfunction that often precedes overt chronic kidney disease by years.

Perhaps most revealing was the age-related pattern. Urinary ACR and PCR climbed steadily with patient age, peaking in the 48-to-57-year group, where ACR reached 20.9 and PCR reached 42.0. Children and adolescents aged 7 to 17 showed values roughly three times lower for ACR and about twice as low for PCR. This gradient supports a cumulative model of renal injury, in which decades of transfusional iron exposure, chronic chelator use, and persistent oxidative stress gradually erode the structural integrity of the kidney’s filtering units.

The pathophysiology is multifactorial. In beta-thalassemia major, ineffective erythropoiesis and chronic hemolysis stem from a critical imbalance between alpha and beta globin chains. Regular transfusions, while life-sustaining, introduce iron faster than the body can safely store it, and suppressed hepcidin worsens systemic iron loading. Excess non-transferrin-bound iron catalyzes the formation of reactive oxygen species, driving lipid peroxidation, mitochondrial dysfunction, and apoptosis in renal tubular and glomerular cells. Histopathologic studies have documented hemosiderin accumulation in proximal tubular epithelial cells, accompanied by tubulointerstitial injury and fibrosis. Persistent oxidative stress also reduces nitric oxide bioavailability, favoring intraglomerular hypertension and protein leakage across the glomerular barrier.

Chronic anemia adds another layer of stress. Reduced oxygen delivery triggers a hyperdynamic circulation and glomerular hyperfiltration, which over time can independently damage the delicate capillary networks of the kidney. Combined with direct iron deposition and inflammatory injury, these mechanisms create a perfect storm for progressive renal dysfunction in long-term survivors.

But the study’s most provocative finding concerns the chelation drugs themselves. When patients were grouped by regimen, those treated with deferasirox, marketed as Exjade, showed the highest levels of proteinuria, followed by those on deferiprone, or Ferriprox, and then deferoxamine, or Desferal. Combination therapy with deferiprone plus deferoxamine produced the lowest urinary values of all. Deferasirox was also the most commonly used chelator in the cohort, accounting for 24 of 44 patients, typically dosed at 20 milligrams per kilogram per day orally.

These observations align with prior literature describing deferasirox-associated proximal tubular dysfunction, reversible rises in serum creatinine, and mild proteinuria. Although deferasirox is primarily excreted through the hepatobiliary route, its systemic iron-chelate complexes may reach the kidney and transiently impair glomerular filtration through interference with prostaglandin-mediated autoregulation. Deferoxamine, administered by subcutaneous infusion, tends to cause fewer renal complications but is linked to auditory, visual, and growth-related adverse events that limit long-term adherence. Deferiprone, an oral bidentate chelator, may induce mild transient proteinuria but is more often associated with agranulocytosis and hepatotoxicity than with nephrotoxicity.

The apparent protective effect of combination therapy is intriguing. Patients receiving both deferiprone and deferoxamine showed the lowest ACR and PCR values while maintaining adequate iron clearance, echoing previous reports suggesting that alternating or combined regimens may optimize the balance between efficacy and safety. If confirmed in larger prospective studies, this pattern could reshape how clinicians individualize chelation strategies for patients at elevated renal risk.

The investigators emphasize that kidney damage in beta-thalassemia is a dynamic, multifactorial process requiring multidisciplinary management. They recommend that regular monitoring of renal function, including assessment of ACR, PCR, and early tubular biomarkers, become an integral part of long-term care for transfusion-dependent patients. Timely detection of subclinical renal dysfunction allows individualized chelation adjustment, optimization of transfusion protocols, and the introduction of nephroprotective measures. As survival continues to improve, preserving kidney function may prove as important as controlling iron itself in this expanding population of long-term thalassemia survivors.

Subject of Research: Renal dysfunction and chelation therapy in beta-thalassemia major

Article Title: Proteinuria and chelator-associated renal dysfunction in β-thalassemia major: insights from a single-center clinical study

Article References: Stoyanova, L., Benkova-Petrova, M., Petrov, A., Abushev, P., Marinov, S., Stoeva, M., & Petrov, Y. (2026). Proteinuria and chelator-associated renal dysfunction in β-thalassemia major: insights from a single-center clinical study. Pharmacia, 73, Article e184854. https://doi.org/10.3897/pharmacia.73.e184854

Image Credits: AI Generated

DOI: 10.3897/pharmacia.73.e184854

Keywords: beta-thalassemia major, proteinuria, albuminuria, iron chelation, deferasirox, deferiprone, deferoxamine, kidney disease, iron overload, renal monitoring, hemoglobinopathy, nephrotoxicity

The review’s central argument is that toxicity prediction is no longer a single-technique discipline but a layered pipeline in which every stage matters. Traditional safety assessment relies on in vitro assays and animal studies, methods the authors describe as resource-intensive, slow, and limited by interspecies physiological differences that can mask human-specific toxicity. Machine learning and deep learning models, by contrast, can extract complex nonlinear patterns from high-dimensional chemical and biological datasets, enabling rapid screening of large compound libraries during early discovery. This shift builds on decades of quantitative structure-activity relationship (QSAR) modeling, which links chemical structure to biological activity, but modern frameworks go further by learning molecular features directly from data rather than relying on hand-crafted rules.

At the foundation of any predictive model lies the question of how to represent a molecule for computation. The review catalogs the dominant approaches: molecular fingerprints such as MACCS structural keys and extended connectivity fingerprints (ECFPs), which encode substructures as bit strings; numerical descriptors calculated by tools like RDKit, Mordred, PaDEL, and the FDA-developed Mold2; and graph-based representations that treat atoms as nodes and bonds as edges. Graph representations have become especially important because they preserve molecular connectivity and allow graph neural networks (GNNs) to learn directly from structure. More recently, pre-trained chemical language models such as ChemBERTa, SMILES-BERT, MolBERT, and MoLFormer generate learned embeddings that capture richer contextual information than fixed fingerprints ever could.

The authors critically compare conventional machine learning algorithms, including random forests, support vector machines, decision trees, k-nearest neighbors, naive Bayes, XGBoost, and LightGBM, against deep learning architectures spanning deep neural networks, convolutional neural networks, GNNs, and transformer-based models. Their conclusion is deliberately unsentimental: no single architecture consistently wins across toxicity tasks. Conventional methods remain competitive on small to medium datasets with informative descriptors, while deep learning approaches benefit from larger datasets and richer representations. Performance differences reported across studies often reflect variations in dataset composition, endpoint definition, class distribution, and validation strategy rather than genuine algorithmic superiority.

One of the review’s most technically significant sections addresses validation rigor, an issue that quietly undermines many published benchmarks. Random train/test splitting can produce optimistic results when structurally similar compounds or close analogues appear in both sets, a problem compounded by data leakage when preprocessing, feature selection, or hyperparameter optimization occurs before the final split. Scaffold-based, cluster-based, temporal, and independent external validation provide far more honest assessments of generalization to novel chemical space. The authors also stress that accuracy alone is insufficient for imbalanced datasets, where toxic compounds are typically rare; they recommend complementary metrics including sensitivity, specificity, F1-score, balanced accuracy, Matthews correlation coefficient, ROC-AUC, and precision-recall AUC, with calibration considered when predicted probabilities inform risk assessment.

The review then organizes its evidence around five organ-specific toxicity domains, drawing on 40 primary studies published between 2021 and 2026. In respiratory toxicity, random forest models using PubChem fingerprints achieved 93.3% accuracy, while a LightGBM model for chemical lung toxicity reached 86.88% accuracy with a ROC-AUC of 0.931. In cardiotoxicity, transformer models using Morgan fingerprints hit a ROC-AUC of 0.93 for hERG channel interactions, and a multi-view GNN framework called CardiacTox achieved a ROC-AUC of 0.938 for CaV1.2 prediction. Nephrotoxicity models reached a ROC-AUC of 0.93 in consensus approaches, while hepatotoxicity studies reported ResNet18 deep neural networks achieving a test ROC-AUC of 0.958 for drug-induced liver injury. Neurotoxicity studies spanned blood-brain barrier permeability, neuronal cytotoxicity, and acetylcholinesterase inhibition, with hybrid deep learning models reaching ROC-AUC values up to 0.964.

These headline numbers, however, come with caveats the authors refuse to bury. Dataset sizes varied enormously, from 2,670 chemicals in a neurotoxicity study to just 83 breast cancer patients in a radiotherapy cardiotoxicity analysis, and KidneyTox_v1.0 relied on only 565 compounds. Class imbalance, uncertain negative labels, limited chemical-space coverage, and incomplete external validation recurred across all five domains. The authors emphasize that a higher ROC-AUC in one organ domain does not indicate a superior modeling approach, because biological endpoints, dataset characteristics, and prediction tasks differ fundamentally across systems. Hepatotoxicity studies typically target drug-induced liver injury, cardiotoxicity studies focus on specific molecular targets like hERG and CaV1.2, and neurotoxicity encompasses heterogeneous endpoints from blood-brain barrier permeability to mammalian neurotoxicity.

Interpretability receives equally critical treatment. Explainable artificial intelligence techniques such as SHAP, attention mechanisms, saliency methods, and structural-alert identification can reveal which molecular features drive a prediction, but the authors draw a sharp line between explaining a model and establishing biological causation. High SHAP values or atom-level attribution maps identify features associated with a prediction without proving those features cause the toxic effect, and attribution results can shift with molecular representation, correlated descriptors, and model architecture. True mechanistic interpretation requires independent evidence from experimental validation, target-interaction data, toxicogenomics, pathway analysis, or adverse outcome pathways. The review also highlights the applicability domain concept, which defines the region of chemical space where a model can be trusted; predictions for compounds far from training data may be unreliable even when internal performance metrics look impressive.

On the regulatory front, the authors argue that strong retrospective performance alone cannot establish suitability for real-world safety assessment. Regulatory confidence depends on independent external validation, clearly defined endpoints, transparent and reproducible workflows, applicability-domain assessment, uncertainty evaluation, and adequate documentation of input data, preprocessing, algorithms, and known limitations. The review catalogs major toxicity databases including SIDER, FAERS, ADReCS, Pneumotox, PubChem, DrugBank, ChEMBL, ICE, LINCS L1000, and LiverTox, noting that clinical adverse-event sources differ fundamentally from experimental toxicity resources and should not be treated as interchangeable. Data quality, endpoint harmonization, and duplicate removal emerge as prerequisites before any model development begins.

Looking forward, the review identifies active learning and federated learning as strategies to overcome limited labeled data, foundation models as a path to stronger molecular representation, and multimodal learning as a way to combine molecular graphs, SMILES strings, structural images, and spectroscopic information. The authors call for improved negative-data construction, assay and endpoint harmonization, causal and mechanistic modeling, and prospective validation to build tools that are reliable, interpretable, and generalizable enough for regulatory applications. Their overarching message is measured but firm: artificial intelligence has demonstrated competitive and sometimes improved predictive performance in toxicology, but it should not be considered inherently superior to conventional methods. The technology’s real value depends on disciplined data curation, rigorous validation, and honest acknowledgment of where predictions can and cannot be trusted.

Subject of Research: Artificial intelligence methods for predicting chemical and drug toxicity across organ-specific endpoints

Article Title: A review on advances in chemical compound toxicology prediction

Article References: Hagag, A., Elsayed, N. A., & Abdelbaky, I. (2026). A review on advances in chemical compound toxicology prediction. Frontiers in Bioinformatics, 6, Article 1974935. https://doi.org/10.3389/fbinf.2026.1974935

Image Credits: AI Generated

DOI: 10.3389/fbinf.2026.1974935

Keywords: toxicity prediction, machine learning, deep learning, graph neural networks, QSAR, hepatotoxicity, cardiotoxicity, nephrotoxicity, neurotoxicity, explainable AI, drug safety, molecular fingerprints

The investigators worked with forty healthy seven-day-old male Shaoxing ducks, a commercially important breed, dividing them into four groups of ten. One group served as untreated controls, one received only Lycium barbarum polysaccharides, or LBPs, at 50 milligrams per kilogram delivered intragastrically, one received cadmium chloride mixed into feed at 50 milligrams per kilogram, and the final group received both. The experiment ran for thirty days under controlled conditions of temperature, humidity, and light. Doses were selected on the basis of previous studies and the team’s own preliminary experiments, and the protocol was approved by the Institutional Animal Care and Use Committee of Yangzhou University.

The consequences of cadmium exposure were quickly apparent. Ducks in the cadmium group lost significant body weight while their kidneys grew heavier relative to body size, a sign of swelling and stress in the organ. Blood chemistry told the same story: levels of creatinine and uric acid, the standard clinical indicators of kidney function, rose sharply. Under the microscope, the damage was striking. Hematoxylin and eosin staining revealed severely disrupted renal architecture, blurred tissue boundaries, atrophied and deformed glomeruli, the kidney’s core filtration units, and hemorrhage in the interstitial spaces. Transmission electron microscopy pushed the view down to the cellular level, exposing nuclear deformation, severe mitochondrial vacuolation, and rupture of the mitochondrial cristae, the folded membranes where cellular energy is produced.

Beneath these visible injuries lay a subtler biochemical sabotage. Using flame atomic absorption spectrometry on digested kidney tissue, the researchers measured the organ’s elemental makeup. Cadmium accumulated dramatically, and as it did, the levels of four essential trace elements, iron, manganese, zinc, and copper, fell. The reason is chemical mimicry. Cadmium shares ionic radius and coordination tendencies with these essential metals, so it can impersonate them in biological systems without performing their functions. The consequences cascade: zinc and copper are components of superoxide dismutase, manganese is the cofactor for the mitochondrial form of that enzyme, and excessive free iron catalyzes the Fenton reaction, generating highly destructive hydroxyl radicals from hydrogen peroxide. Deprived of their raw materials, the antioxidant enzymes falter precisely as the toxic metal floods the system.

The oxidative stress measurements confirmed this collapse. Malondialdehyde, a marker of lipid peroxidation, and hydrogen peroxide rose markedly, while the activities of superoxide dismutase, catalase, and total antioxidant capacity fell, along with the content of glutathione, the cell’s master antioxidant. The team then traced the failure to a master molecular switch. Nrf2 is a transcription factor that, under stress, detaches from its cytoplasmic anchor Keap1, moves into the nucleus, and switches on antioxidant response element-dependent genes. In cadmium-exposed ducks, the protein levels of Nrf2 and its downstream targets, HO-1, NQO1, GCLC, GCLM, and GSR, all dropped. Because GCLC, GCLM, and GSR govern the synthesis and regeneration of glutathione, their suppression crippled the cell’s capacity to manufacture and recycle its most important antioxidant defense.

This is where the wolfberry compound intervened. When ducks received both cadmium and LBP, the pattern reversed. Cadmium burden in the kidney fell, iron, manganese, zinc, and copper levels recovered, the antioxidant enzymes rebounded, and the Nrf2 pathway proteins climbed back toward normal. The researchers attribute the effect to the intrinsic antioxidant, anti-inflammatory, and lipid-lowering properties of LBPs, the major bioactive components of the goji berry, which have been valued in traditional Chinese medicine and are increasingly supported by modern pharmacology. Previous work has shown these polysaccharides acting through multiple signaling pathways with anti-aging, hypoglycemic, immunomodulatory, and neuroprotective effects, and dietary LBP supplementation has already been shown to enhance antioxidant defenses and growth performance in broiler chickens.

But the study’s most novel contribution extends beyond the classic antioxidant story. Cadmium, the team found, triggered a profound disruption of lipid metabolism in the kidney. Total cholesterol and triglyceride levels rose in both serum and kidney tissue, and lipid droplets accumulated visibly under the electron microscope. At the protein level, cadmium upregulated the lipogenic machinery, including FAS, SREBP1, ACC1, ACC2, and PPARγ, the transcription factor that drives adipocyte differentiation and lipid storage. Simultaneously it downregulated the fat-burning side of the ledger: ATGL and HSL, the rate-limiting enzymes that break triglycerides down step by step, along with PPARα, the master regulator of fatty acid oxidation, and CPT2 and LPL, which shuttle fatty acids into mitochondrial beta-oxidation and clear circulating triglyceride-rich lipoproteins respectively. LBP co-treatment corrected both directions of this imbalance, restoring fat breakdown while suppressing excess fat synthesis.

That metabolic rescue rippled into the immune system. The kidney is not merely a filter but an active immune organ, populated by macrophages and T cells that maintain its internal environment. Macrophages polarize into pro-inflammatory M1 cells, marked by iNOS and CD68, or anti-inflammatory M2 cells, marked by Arg1, while T helper cells divide similarly into Th1 and Th2 lineages, tracked by T-bet and GATA3. Cadmium exposure pushed the balance hard toward inflammation: iNOS, CD68, T-bet, and CD4 rose while Arg1 and GATA3 fell, and the inflammatory cytokines IL-18, IL-1β, and TNF-α climbed in kidney tissue. LBP supplementation reversed the polarization markers and substantially lowered all three cytokines. The finding fits into a wider picture in which oxidative stress and disordered lipid metabolism feed chronic inflammation, exhaust immune cells, and ultimately produce immunosuppression, a vicious cycle that the polysaccharide appears to interrupt at multiple points.

The authors are candid about their limits. All results come from phenotypic observations and targeted protein measurements in a live duck model, and without Nrf2 inhibitor or gene-interference experiments the Nrf2 findings indicate correlation rather than proven causality. No transcriptomic or metabolomic analyses were performed, so the global molecular landscape remains unmapped. Even so, the study offers molecular evidence for a natural, widely available feed additive that could protect livestock from heavy metal nephrotoxicity, with implications that reach the human dinner plate, since duck meat and offal enter the food supply chain. As cadmium pollution intensifies with industrialization, a humble berry extract that quiets oxidative stress, restores trace element balance, mends lipid metabolism, and calms inflamed immune tissue is a finding worth watching closely.

Subject of Research: Protective effects of Lycium barbarum polysaccharides against cadmium-induced kidney injury in ducks via the Nrf2 antioxidant pathway

Article Title: Lycium barbarum polysaccharides alleviate cadmium-induced renal injury in ducks by activating the Nrf2-mediated antioxidant defense pathway

Article References: Wang, X., Liu, D., Ma, Y., Zou, H., Tong, X., Zhu, J., Song, R., Zhao, H., & Liu, Z. (2026). Lycium barbarum polysaccharides alleviate cadmium-induced renal injury in ducks by activating the Nrf2-mediated antioxidant defense pathway. Frontiers in Veterinary Science, 13, Article 1998071. https://doi.org/10.3389/fvets.2026.1998071

Image Credits: AI Generated

DOI: 10.3389/fvets.2026.1998071

Keywords: cadmium, nephrotoxicity, Lycium barbarum polysaccharides, Nrf2, oxidative stress, ducks, lipid metabolism, kidney injury, trace elements, antioxidant defense, immune balance, wolfberry

Now a large international team spanning Spain and Slovenia has delivered what the field has lacked for decades: a single, rigorously controlled comparison of 26 aminoglycosides and their molecular components, tested side by side under identical conditions. The study, led by Nadia García-Mateo, Hrvoje Petković, and M. Beatriz Durán-Alonso, appears in Frontiers in Pharmacology. The researchers evaluated antibacterial potency against 61 well-characterized clinical isolates from the ESKAPE panel, a collection that includes carbapenemase-producing enterobacteria, multidrug-resistant Pseudomonas aeruginosa, and multidrug-resistant Acinetobacter baumannii, the pathogens that pose the gravest threat to hospitalized patients. In parallel, they measured toxicity using two otic cell lines derived from the organ of Corti, the OC-k3 and HEI-OC1 lines, plus the porcine kidney cell line LLC-PK1, and they extended key findings to organotypic cochlear explants from newborn mice.

The antibacterial screen produced clear winners. Eight commercial aminoglycosides demonstrated strong activity against the multidrug-resistant isolates: amikacin, apramycin, gentamicin, isepamicin, neomycin, paromomycin, sisomicin, and tobramycin. Apramycin and amikacin emerged as the most broadly active agents, each inhibiting 90 percent of the isolates at 32 milligrams per liter, while several competitors failed to reach meaningful inhibition even at 64 milligrams per liter. The lowest MIC50 value, just 2 milligrams per liter, belonged to the neomycin preparations, and the neomycin B congener showed the strongest individual performance. Four isolated congeners, gentamicin C1a, C2, C2a, and neomycin B, also proved potent, alongside the intriguing observation that kanamycin B slightly outperformed both its parental mixture and the kanamycin A congener.

Toxicity testing revealed a far more complicated picture. Hygromycin B was the most toxic compound on both otic cell lines even at low concentrations and short exposures, while kasugamycin and streptomycin also damaged both lines significantly. Kanamycin, ribostamycin, and tobramycin showed no significant loss of cell viability in the MTT assays, and apramycin, isepamicin, paromamine, and paromomycin were toxic only at the higher 5 millimolar concentration. Notably, the two otic cell lines sometimes disagreed: sisomicin selectively damaged OC-k3 cells while sparing HEI-OC1 cultures, and amikacin’s apparent toxicity to HEI-OC1 cells in the MTT assay vanished when the researchers used an ATP-based viability test instead. The cochlear explant data suggested the ATP assay and the OC-k3 line together offer the more reliable predictions.

The most striking results came from dissecting gentamicin itself. Commercial gentamicin is not a single molecule but a fermentation-derived mixture of structurally related congeners, dominated by the C-subtypes C1, C1a, C2, C2a, and C2b, with minor quantities of A, B, B1, and X2 variants. For the first time, the team directly compared seven of the most common gentamicin congeners for both antibacterial activity and toxicity. At 2 millimolar concentrations, gentamicin B1 and gentamicin X2 devastated the otic cell lines, dropping viability to single digits or low double digits on the OC-k3 line, an outcome dramatically worse than the gentamicin mixture itself, which left more than 80 percent of cells viable. These congeners also showed very weak antibacterial activity, with gentamicin X2 among the poorest inhibitors of the bacterial ribosome, and gentamicin X2 proved the most nephrotoxic of all the gentamicin congeners on the LLC-PK1 kidney cell line.

The implications for clinical practice are sobering. Because the congener composition of commercial gentamicin varies across manufacturers and even between fermentation batches, even small amounts of B1 or X2 contamination could meaningfully alter the overall toxicity of a given formulation. European regulators already require gentamicin B1 to be identified as a specified impurity, but gentamicin X2 has not received the same scrutiny. The researchers point to structural clues: gentamicin B1 and X2 differ from the C-group at specific positions on the first sugar ring, echoing the arrangement found in sisomicin, which also showed high ototoxicity. They suggest that the exact geometry of this ring likely modulates how these compounds interact with cellular targets, though the precise mechanisms behind the extreme toxicity of B1 and X2 remain unresolved and are probably multifactorial.

The ex vivo work on murine cochlear explants sharpened the picture further. Amikacin and isepamicin left both outer and inner hair cells essentially unscathed, with survival rates slightly above untreated controls, while tobramycin and sisomicin destroyed most hair cells in the cultured cochlear epithelium. Interestingly, applying plain neomycin caused a significant loss of outer hair cells that was not observed with the neomycin B congener, even though the two performed identically in cell-line assays. The explants also reproduced a well-known vulnerability gradient, with hair cells at the basal end of the cochlea, which encode high frequencies, dying first. The concentration used, 0.1 millimolar, sits within the range of peak plasma levels measured in patients treated with gentamicin, isepamicin, or amikacin, lending physiological relevance to the model.

On the kidney side, amikacin earned a singular distinction: it was the only strongly antibacterial aminoglycoside that produced no significant nephrotoxicity in the LLC-PK1 proximal tubule cell line. This aligns with clinical experience, in which amikacin is generally regarded as less nephrotoxic than gentamicin. Surprisingly, every isolated gentamicin congener, including the least toxic C1a and C2a variants, proved more nephrotoxic than the commercial mixture, and neomycin B was significantly more toxic to kidney cells than its parental mixture. Aminoglycosides accumulate in kidney lysosomes and bind negatively charged phospholipids, but the researchers could not find a simple correlation between charge or structure and the observed nephrotoxicity ranking, underscoring how much remains unknown about these mechanisms.

The authors are careful about the limits of their work. All toxicity measurements were performed on cell lines and explants at concentrations well above those typically found in patients’ plasma, which vary with dosing regimens, tissue distribution, and the infectious agent, so confirmation in living animal models, such as guinea pig cochlear studies, will be essential before the findings reshape clinical practice. Still, the methodological validation is strong: this same in vitro pipeline previously flagged gentamicin C1a as less ototoxic than the mixture, a result later confirmed in vivo, and the consistent toxicity signals from gentamicin B1 and X2 across every assay lend credibility to the alarm. As biosynthetic engineering and scalable purification now make it economically feasible to isolate individual congeners at industrial scale, the study suggests a dual path forward: refining existing formulations to exclude the most dangerous minor components, and mining natural aminoglycoside diversity, exemplified by semisynthetic successes like plazomicin and etimicin, for new drug leads that separate bacterial killing from cochlear and renal destruction.

Subject of Research: Comparative antibacterial activity, ototoxicity and nephrotoxicity of aminoglycoside antibiotics and their congeners

Article Title: Comparative evaluation of the antibacterial activity and toxicity of a large panel of aminoglycoside antibiotics

Article References: García-Mateo, N., López-Hernández, I., Čusak, A., Fernández-Martínez, M., Šala, M., Schimmang, T., Kosec, G., Fujs, Š., Martínez-Martínez, L., Petković, H., & Durán-Alonso, M. B. (2026). Comparative evaluation of the antibacterial activity and toxicity of a large panel of aminoglycoside antibiotics. Frontiers in Pharmacology, 17, Article 1930759. https://doi.org/10.3389/fphar.2026.1930759

Image Credits: AI Generated

DOI: 10.3389/fphar.2026.1930759

Keywords: aminoglycosides, gentamicin, antibiotic resistance, ototoxicity, nephrotoxicity, ESKAPE pathogens, congeners, cochlear explants, kidney cell line, MIC90, multidrug resistance, antimicrobial susceptibility

The research, led by Yoshitaka Saito of the Department of Clinical Pharmaceutics and Therapeutics at Hokkaido University of Science together with colleagues at Hokkaido University Hospital, examined 244 patients with thoracic cancers who were treated with cisplatin-based chemotherapy at doses of at least 75 milligrams per square metre of body surface area. Crucially, all of these patients received what is known as the short hydration method, a streamlined fluid regimen that has rapidly become the standard of care for preventing cisplatin-induced nephrotoxicity, the clinical term for chemotherapy-driven kidney injury.

To understand why this finding matters, it helps to grasp how cisplatin attacks the kidney. After intravenous infusion, the drug is filtered by the kidneys and accumulates preferentially in the proximal tubules, the microscopic workhorses responsible for reabsorbing water, salts and nutrients. There, platinum complexes bind to cellular proteins and DNA, triggering oxidative stress, mitochondrial dysfunction and ultimately cell death. The result is a rise in serum creatinine, a waste product that healthy kidneys efficiently clear, and a corresponding drop in creatinine clearance, the standard laboratory measure of how well the kidneys are filtering blood. Clinicians track both values closely during every cycle of cisplatin therapy.

For decades, the answer to this toxicity was aggressive hydration: patients were flooded with litres of saline over many hours, sometimes with forced diuresis using diuretics, to flush the drug through the renal tubules before it could do damage. That approach works, but it is burdensome, often requiring hospital admission for a day or more per cycle. The short hydration method, by contrast, delivers a carefully calibrated volume of fluid over a much shorter window, typically a few hours around the cisplatin infusion, allowing outpatient treatment without sacrificing protection. It represents the most advanced prophylaxis currently available against cisplatin-induced nephrotoxicity.

That is precisely what makes the new result so consequential. Earlier studies had hinted that hypertension might be a risk factor for cisplatin kidney injury, but those analyses were conducted in settings where kidney protection was less rigorous, leaving open the possibility that the association was really a proxy for inadequate hydration rather than a genuine biological vulnerability. By studying a cohort that uniformly received short hydration, the Hokkaido team could test whether hypertension remains a risk factor even under optimal modern prophylaxis. The answer, according to their data, is yes.

Patients were divided into two groups: those with baseline hypertension requiring pharmacological treatment before starting chemotherapy, and a control group without hypertension. Across all treatment cycles, cisplatin-induced nephrotoxicity developed in 18.0 percent of the hypertensive patients, compared with just 6.0 percent of the controls, a threefold difference that reached statistical significance with a P value of 0.008. This comparison served as the study’s primary endpoint, and it was met decisively.

The damage was not limited to a binary yes-or-no outcome. When the researchers examined continuous measures of kidney function, they found that the maximum changes in serum creatinine from baseline were significantly worse in the hypertensive group, with a P value of 0.009, and the maximum changes in creatinine clearance were also notably worse, at P equal to 0.0496. In other words, hypertensive patients did not merely cross a laboratory threshold more often; their kidneys measurably declined further during treatment than those of their normotensive counterparts.

Because this was a retrospective, real-world study rather than a randomised trial, the investigators took careful steps to rule out confounding. They applied propensity score matching, a statistical technique that pairs hypertensive and non-hypertensive patients who share similar baseline characteristics, effectively mimicking the balance of a randomised comparison. The matched analysis reproduced the original findings. The team also repeated the analysis after excluding patients who were concurrently receiving bevacizumab, an anti-cancer antibody known to affect blood pressure and kidney function, and the results remained consistent. Together, these sensitivity analyses strengthen the conclusion that hypertension itself, and not some co-administered drug or demographic imbalance, drives the excess risk.

The biological plausibility of the finding is compelling. Chronic hypertension damages the small vessels of the kidney, causing thickening and narrowing of the arterioles that supply the glomeruli and tubules, a process known as nephrosclerosis. A kidney already operating with compromised blood flow has less physiological reserve to withstand a second insult such as cisplatin. Reduced renal perfusion may also prolong the residence time of the drug within tubular cells, increasing local exposure. In addition, many antihypertensive medications, particularly agents acting on the renin-angiotensin system, alter renal haemodynamics in ways that could interact with the dehydration and electrolyte shifts that accompany cisplatin therapy, although the study did not dissect these mechanisms in detail.

For oncologists and clinical pharmacists, the practical message is clear: baseline hypertension should be treated as a red flag when planning cisplatin-based chemotherapy under short hydration. Patients with treated hypertension may warrant closer monitoring of serum creatinine and creatinine clearance across cycles, earlier nephrology consultation, or consideration of alternative platinum agents such as carboplatin in selected cases. The authors, whose work was funded by JSPS KAKENHI grant number 26K09780 and approved by the Ethical Review Board of Hokkaido University Hospital, note that their findings identify patients with baseline hypertension as being at high risk of developing cisplatin-induced nephrotoxicity even with short hydration. As short hydration continues to spread worldwide as the default kidney-protection strategy, this study serves as a timely reminder that the best protocols still leave vulnerable patients exposed, and that personalised risk assessment, beginning with a simple blood pressure reading taken before the first infusion, may be one of the cheapest and most effective tools in modern cancer care.

Subject of Research: The association between baseline hypertension and cisplatin-induced nephrotoxicity in cancer patients receiving short hydration

Article Title: Association between baseline hypertension and cisplatin-induced nephrotoxicity in patients receiving short hydration

Article References: Saito, Y., Kobayashi, M., Sakamoto, T., Takekuma, Y., Sakakibara-Konishi, J., Shimizu, Y., Kinoshita, I., & Sugawara, M. (2026). Association between baseline hypertension and cisplatin-induced nephrotoxicity in patients receiving short hydration. Scientific Reports. https://doi.org/10.1038/s41598-026-75098-8

Image Credits: AI Generated

DOI: 10.1038/s41598-026-75098-8

Keywords: cisplatin, nephrotoxicity, hypertension, short hydration, kidney injury, chemotherapy, oncology, serum creatinine, creatinine clearance, propensity score matching, thoracic cancer, clinical pharmacy

The research team, led by Jane Namukobe and Moses Murungi of Makerere University’s Department of Chemistry, set out to fill that gap using Wistar albino rats, the standard mammalian workhorse of preclinical safety testing. Plant material was collected between April and June from Mpanga Central Forest Reserve and Mabira Forest, authenticated at the Makerere University herbarium, and processed into two very different preparations. Aqueous extracts, made by warm maceration in water at 40 degrees Celsius, were reserved for oral toxicity testing because that is how the plants are traditionally consumed. Organic extracts, prepared with a one-to-one mixture of dichloromethane and methanol, were used for dermal testing because topical remedies are typically blended with vehicles such as petroleum jelly. This design choice matters: what is safe to swallow and what is safe to smear are separate questions, and the study treated them as such.

The testing regime followed Organisation for Economic Co-operation and Development guidelines with slight modifications. For acute toxicity, rats received single oral doses of 2000 or 5000 milligrams per kilogram of body weight, or single dermal doses of 8000 or 10,000 milligrams per kilogram. For repeated-dose studies, animals were dosed daily with 200 or 400 milligrams per kilogram orally for 28 days, or 2000 and 5000 milligrams per kilogram dermally for 21 days. Throughout the studies, researchers recorded clinical observations, tracked body weights weekly, and then euthanised the animals for a full workup: blood chemistry covering liver and kidney markers, organ weights, and histopathological examination of the liver, kidneys, heart, stomach and small intestines. Skin irritation was scored with the Draize system, and sensitisation was probed with a modified mouse ear swelling test using tape stripping to heighten immune responses.

The headline finding is deceptively reassuring: no animal died, and no acute toxicity was observed for any extract by either route. Plectranthus caespitosus emerged as the standout performer. Its aqueous leaf extract produced no clinical signs, no changes in body or organ weights, no biochemical abnormalities and clean histopathology even after 28 days of repeated oral dosing. Dermal application caused neither irritation nor sensitisation. This is the first toxicity study ever conducted on this species, and the result aligns with related Plectranthus species such as P. neochilus and P. amboinicus, which have similarly shown wide safety margins in rodents. For a plant already used against skin allergies, the data make a strong case for pushing its active compounds toward antibacterial and skincare applications.

Erlangea tomentosa tells a more troubling story. A single oral dose of its leaf extract, even at the maximum 5000 milligrams per kilogram, caused only transient effects: elevated respiration, reduced mobility and increased urination that resolved within 24 hours. But repeated oral exposure at 400 milligrams per kilogram produced moderate to severe multifocal necrotizing hepatitis, literally patches of dying liver tissue, along with hyperplasia of goblet cells in the small intestine and sloughing of intestinal villi. Remarkably, the liver damage was invisible in the blood chemistry: bilirubin, creatinine and alkaline phosphatase levels remained statistically indistinguishable from controls, and the ALT-to-ALP ratio stayed below 10. The researchers attribute this mismatch to the liver’s known capacity for rapid biochemical recovery once a necrotizing agent is withdrawn, a phenomenon documented in drug-induced liver injury literature. Histopathology, in other words, caught what blood tests missed.

Repeated dermal exposure to E. tomentosa’s organic extract added another layer of concern. Although the extract was classified as a non-irritant with a primary irritation index below 1.5 and provoked no sensitisation in the ear swelling test, animals dosed repeatedly at 200 and 400 milligrams per kilogram developed necrotizing liver hepatitis. Blood chemistry showed significantly reduced ALT, ALP and urea alongside elevated bilirubin, a pattern the authors note can indicate cholestasis or impaired bile clearance, though reduced transaminases can also reflect hepatoprotective effects. The team points to historical reports of the plant’s toxicity to rabbits, in which fatal doses caused abdominal ascites and congestion of the liver, kidneys and lungs. Phytochemical screening offers a plausible mechanism: the extract contains tannins and anthraquinones, both of which have been shown to induce hepatic necrosis at sufficient concentrations, while its saponins may partially counteract that damage.

Psorospermum febrifugum, the stem bark remedy used across tropical Africa for skin diseases and wounds, occupied a middle ground. Acute oral exposure at 5000 milligrams per kilogram increased liver weight and temporarily raised respiration, urination and mobility, all resolving within a day. Repeated oral dosing at 400 milligrams per kilogram reduced both liver and kidney weights and produced a roughly three-fold increase in the AST-to-ALT ratio, a signature that typically signals hepatocellular injury or stress, together with elevated total and direct bilirubin. Yet histopathology of the liver showed no significant lesions, a discrepancy the researchers interpret as either adaptive organ remodelling, the liver’s regenerative capacity masking early damage, or the slow onset of chronic toxicity. The plant’s tannins and saponins, with their opposing effects on liver tissue, may explain why the tissue looked healthy even as enzyme levels flagged distress.

The dermal picture for P. febrifugum was arguably the most clinically significant. Topical application caused no skin irritation or sensitisation, but repeated exposure at 2000, 5000 and 8000 milligrams per kilogram significantly increased bilirubin, creatinine and AST while decreasing urea and ALP, a combination the authors link to drug-induced cholestasis and impaired kidney function. Acute exposure at 8000 milligrams per kilogram also disturbed sodium and potassium levels. The organic extract is rich in steroids, compounds known to induce or aggravate acute kidney injury, chronic kidney disease and glomerular toxicity, providing a chemical suspect for the renal findings. These results are notable because they emerged from skin application alone, demonstrating that a plant remedy can damage internal organs without ever touching the digestive tract, a fact with real implications for how topical herbal medicines are regulated.

The broader lesson of the study is that safety cannot be inferred from tradition, efficacy, or even a single toxicity endpoint. All three plants passed the most visible tests, no mortality, no acute toxicity, no skin irritation, yet repeated exposure revealed liver and kidney liabilities in two of them, and in E. tomentosa the worst damage appeared only under the microscope. The authors recommend that P. caespitosus advance toward development of antibacterial products and skincare formulations, while E. tomentosa and P. febrifugum require further work to isolate and characterise their toxic constituents, with tannins flagged as the most likely culprits. As antimicrobial resistance drives patients and researchers alike back toward plant-based remedies, this study stands as a cautionary benchmark: the dose, the route and the duration of exposure all matter, and a herbal medicine is only as safe as its most thorough toxicological evaluation.

Subject of Research: Oral and dermal toxicity assessment of three Ugandan ethno-medicinal plants used to treat bacterial skin infections, tested in Wistar albino rats

Article Title: Assessment of oral and dermal toxicity of selected ethno-medicinal plants used to treat bacterial skin infections in Uganda using Wistar albino rats

Article References: Namukobe, J., Murungi, M., Asiimwe, S., Tayebwa, D. S., Mukwaya, J., Aturagaba, G., Sekandi, P., & Byamukama, R. (2024). Assessment of oral and dermal toxicity of selected ethno-medicinal plants used to treat bacterial skin infections in Uganda using Wistar albino rats. Discover Toxicology, 1(1), Article 12. https://doi.org/10.1007/s44339-024-00015-5

Image Credits: AI Generated

DOI: 10.1007/s44339-024-00015-5

Keywords: ethnomedicine, toxicology, Uganda, medicinal plants, hepatotoxicity, nephrotoxicity, Wistar rats, skin infections, phytochemistry, OECD guidelines, dermal toxicity, antimicrobial resistance

The study, published in Supportive Care in Cancer, was led by Syed E. Pasha of the University of Houston and HCA Houston Healthcare Kingwood, together with colleagues from MountainView Hospital and Texas Oncology centers. The team set out to answer a deceptively simple question: do patients with diffuse large B-cell lymphoma, an aggressive and rapidly treated lymphoma, experience different kidney outcomes than patients with chronic lymphocytic leukemia or small lymphocytic lymphoma, an indolent disease that often smolders for years before requiring therapy? Comparative renal data between these two populations had been scarce, despite the fact that both cancers are common and both are treated with regimens known to stress the kidneys.

To answer the question, the researchers turned to the TriNetX Global Collaborative Network, a federated database drawing de-identified electronic health records from 143 healthcare organizations. Data were extracted on May 9, 2025. Adults with a coded diagnosis of diffuse large B-cell lymphoma or of chronic lymphocytic leukemia and small lymphocytic lymphoma were identified using standard ICD-10 diagnostic codes. To make the two groups as comparable as possible, the investigators used a statistical technique called propensity score matching, pairing each patient with a counterpart from the other cohort who shared the same age, sex, race, and ethnicity. This one-to-one matching produced two demographically indistinguishable groups, with standardized mean differences between them ranging from just 0.001 to 0.038, values so small they indicate near-perfect balance.

The final analysis included 19,438 patients, evenly split with 9,719 in each cohort. Kidney outcomes were tracked from the first day after the index diagnosis through each patient’s available follow-up, an open-ended window rather than a fixed observation period. The researchers looked for two distinct conditions: acute kidney injury, the sudden loss of renal function that can develop within hours or days, and chronic kidney disease, the gradual, often irreversible decline in filtration capacity that unfolds over months and years. Statistical comparisons relied on two-proportion z-tests, generating risk differences, risk ratios, and odds ratios with 95 percent confidence intervals.

The results revealed a fascinating split. Acute kidney injury struck 22.9 percent of the diffuse large B-cell lymphoma patients compared with 20.9 percent of the chronic lymphocytic leukemia and small lymphocytic lymphoma group. That two-percentage-point gap was statistically significant, with a risk ratio of 1.094 and an odds ratio of 1.122, and a p-value of 0.001. In other words, the aggressive lymphoma carried a modestly higher risk of sudden renal failure. The pattern flipped for chronic kidney disease. Here, 18.8 percent of the lymphoma patients developed the condition versus 22.7 percent of the leukemia and indolent lymphoma group, a negative risk difference of 4.0 percent, with a risk ratio of 0.825 and an odds ratio of 0.785, and a p-value below 0.001. The slower cancer, paradoxically, left a heavier mark on long-term kidney health.

The biology behind this divergence makes intuitive sense when one considers how each disease and its treatment unfold. Diffuse large B-cell lymphoma grows quickly and is typically treated with intensive chemoimmunotherapy, most commonly regimens built around rituximab combined with cyclophosphamide, doxorubicin, vincristine, and prednisone. Rapid tumor cell death can trigger tumor lysis syndrome, a cascade in which potassium, phosphate, and uric acid flood the bloodstream and crystallize in renal tubules, a well-documented cause of acute kidney injury. High-dose chemotherapy agents themselves, along with the infections, dehydration, and intensive care admissions that can accompany aggressive treatment, all add to the short-term renal assault.

Chronic lymphocytic leukemia, by contrast, is a disease of accumulation and time. Patients frequently live with it for a decade or more, and even those who never require treatment face aging kidneys, cumulative exposure to nephrotoxic drugs across successive lines of therapy, and immune complications. Modern targeted therapies such as ibrutinib, which has transformed outcomes in this disease, carry their own renal baggage; real-world analyses of hundreds of treated patients have documented a spectrum of toxicities that can include hypertension and kidney dysfunction. Over years of surveillance, treatment, and comorbidity, the indolent disease’s patient population gradually accrues the chronic damage that shows up as chronic kidney disease. The new data, showing a 22.7 percent chronic kidney disease rate in this group, align with that clinical picture.

The authors are careful to frame their findings as hypothesis-generating rather than definitive. The matching algorithm balanced demographics but did not capture comorbidities such as diabetes, hypertension, or pre-existing renal impairment, all of which are powerful drivers of kidney disease and may differ systematically between the two cancer populations. Follow-up duration was also not matched, and the open-ended observation window means patients in one group may simply have been watched longer. Absolute differences between the cohorts were small, in the range of two to four percentage points, and the study relied on administrative diagnostic codes rather than laboratory-verified kidney function measures. The research was supported in whole or in part by HCA Healthcare, though the authors declare no competing interests, and the views expressed are those of the authors alone.

Even with those caveats, the study carries practical weight for the growing field of onco-nephrology, the discipline that sits at the intersection of cancer care and kidney medicine. Acute kidney injury in patients with hematological malignancies is known to worsen remission rates and survival, and chronic kidney disease constrains which cancer therapies a patient can safely receive, since many agents require dose adjustment or are contraindicated in renal impairment. If the pattern observed here holds up, supportive care pathways might reasonably emphasize vigilant short-term renal monitoring, hydration protocols, and uric acid management around the intensive treatment phases of aggressive lymphoma, while building durable long-term nephrology follow-up into the care of chronic lymphocytic leukemia survivors, whose kidneys face a marathon rather than a sprint.

What comes next is confirmation. The authors themselves call for comorbidity-adjusted analyses and prospective studies to verify whether the renal risk differences survive when diabetes, hypertension, baseline kidney function, and treatment exposure are properly accounted for. Until then, the message for clinicians is one of heightened awareness rather than alarm: both cancer subtypes deserve deliberate attention to the kidneys, but the timing and nature of that attention may need to differ. A single dataset of nearly twenty thousand matched patients cannot settle every question, but it has usefully drawn a map, pointing oncologists toward the moments in each disease trajectory where the kidneys are most vulnerable, and reminding the wider medical community that surviving cancer is only half the story, with protecting the organs that carry patients through survivorship being the other half.

Subject of Research: Comparative renal outcomes of acute kidney injury and chronic kidney disease in diffuse large B-cell lymphoma versus chronic lymphocytic leukemia/small lymphocytic lymphoma

Article Title: Renal risk differences between diffuse large B-cell lymphoma and chronic lymphocytic leukemia/small lymphocytic lymphoma: implications for supportive oncology care

Article References: Pasha, S. E., Khalique, S., Haddadin, R., Mazharuddin, S., & Khoury, P. N. (2026). Renal risk differences between diffuse large B-cell lymphoma and chronic lymphocytic leukemia/small lymphocytic lymphoma: implications for supportive oncology care. Supportive Care in Cancer, 34(10), Article 1029. https://doi.org/10.1007/s00520-026-11256-y

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11256-y

Keywords: diffuse large B-cell lymphoma, chronic lymphocytic leukemia, small lymphocytic lymphoma, acute kidney injury, chronic kidney disease, supportive oncology, onco-nephrology, nephrotoxicity, TriNetX, propensity score matching, real-world data, tumor lysis syndrome