Researchers searched PubMed, Embase, Web of Science and the Cochrane Library for studies evaluating blood-based biomarkers in pancreatic cancer diagnosis. They selected the 20 most commonly reported biomarkers and extracted sensitivity, specificity and other diagnostic information from 185 eligible studies. The review then used a Bayesian analysis-of-variance model to pool and compare the biomarkers’ diagnostic measures.

The statistical model generated both overall rankings and separate assessments of sensitivity and specificity. IMP3, S100P and CA195 occupied the top three positions when the researchers calculated the dominant index. IMP3 also appeared among the strongest performers for specificity, while S100P received the highest sensitivity ranking in the frequentist analysis.

The frequentist analysis, which used SUCRA to summarize the ranking of each biomarker, produced a different leading group for sensitivity. S100P ranked first, followed by Span-1 and MIC-1. For specificity, p53 ranked first, with IMP3 and CA72-4 completing the top three. The variation across ranking methods shows why the results should be understood as a comparative summary rather than a single definitive test order.

The study conducted subgroup analyses across different study designs, control groups and geographic regions. The authors reported that the results in these analyses were consistent with the overall findings. The review did not establish that any biomarker should replace CA19-9 in routine practice, and the authors said that IMP3 and S100P warrant further investigation in prospective studies to confirm their clinical utility.

CA19-9 remains the most widely used serum biomarker for pancreatic cancer, but the review noted that its sensitivity and specificity are suboptimal, limiting its clinical utility. The authors also emphasized that the analysis compares published diagnostic performance across multiple biomarkers. As a result, the ranked findings identify markers for additional evaluation but do not by themselves demonstrate improved patient outcomes or a complete clinical testing strategy.

Overall, the review points to IMP3 and S100P as promising blood-based markers with high diagnostic accuracy in the comparative analysis. CA195 also ranked strongly in the Bayesian assessment, while p53, Span-1, MIC-1 and CA72-4 led particular frequentist rankings. These findings support prospective investigation of the most consistently performing candidates, while the need for clinical confirmation remains central to the study’s conclusion.

Subject of Research: Comparative diagnostic accuracy of blood-based biomarkers for pancreatic cancer

Article Title: Comparative diagnostic accuracy of twenty blood-based biomarkers for pancreatic cancer: a network meta-analysis

Article References: Lu, T.-L., Gao, J., Liu, R.-H., & Chen, C.-W. (2026). Comparative diagnostic accuracy of twenty blood-based biomarkers for pancreatic cancer: a network meta-analysis. BMC Cancer. https://doi.org/10.1186/s12885-026-17140-6

Image Credits: AI Generated

DOI: 10.1186/s12885-026-17140-6

Keywords: pancreatic cancer, blood biomarkers, network meta-analysis, diagnostic accuracy, IMP3, S100P, CA195, CA19-9, sensitivity, specificity, systematic review, Comparative

The review compared nine preventive interventions: three durations of scalp hypothermia, scalp compression, four pharmacological agents, and usual care. The analysis included trials reporting hair-preservation outcomes such as CTCAE grade below 2, hair loss below 30% to below 50%, or WHO grade 0–2. The researchers searched PubMed, Embase, CENTRAL, and EBSCO Open Dissertations from database inception through March 2025. Risk ratios were pooled with a random-effects model, interventions were ranked using the surface under the cumulative ranking curve (SUCRA), and certainty was evaluated with CINeMA.

For prolonged scalp hypothermia, defined in the review as more than 90 minutes after chemotherapy, the pooled risk ratio for improved hair preservation was 24.66 compared with usual care (95% confidence interval, 7.07–86.05). The corresponding risk ratio was 2.13 compared with traditional scalp hypothermia lasting 46–90 minutes (95% CI, 1.15–3.93). Prolonged scalp hypothermia also produced better hair-preservation results than the four pharmacological agents, with risk ratios ranging from 14.06 to 23.83. The evidence for these comparisons was graded as low certainty. A comparison with short-duration scalp hypothermia, lasting 0–45 minutes, favored prolonged treatment numerically, but was not statistically significant: RR 2.06 (95% CI, 0.78–5.39).

The SUCRA ranking placed prolonged scalp hypothermia highest at 97.0%, followed by short-duration scalp hypothermia at 78.6% and traditional scalp hypothermia at 76.8%. These rankings indicate the ordering generated by the network analysis, but the review’s authors noted that most included studies had a high risk of bias and that the certainty of evidence ranged from low to very low. The abstract therefore supports the conclusion that scalp hypothermia could prevent chemotherapy-induced alopecia, while the difference between prolonged and short treatment was not statistically established. The review was registered in PROSPERO as CRD420251103932.

Subject of Research: Comparative effectiveness of interventions for preventing chemotherapy-induced alopecia

Article Title: Comparative efficacy of interventions for preventing chemotherapy-induced alopecia: a systematic review and network meta-analysis

Article References: Boonpattharatthiti, K., Lawin, S., Limhengsin, P., Dhippayom, T., & Danwilai, K. (2026). Comparative efficacy of interventions for preventing chemotherapy-induced alopecia: a systematic review and network meta-analysis. Supportive Care in Cancer, 34(11), Article 1123. https://doi.org/10.1007/s00520-026-11354-x

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11354-x

Keywords: chemotherapy-induced alopecia, scalp hypothermia, network meta-analysis, randomized controlled trials, hair preservation, systematic review, risk of bias, evidence certainty, Comparative, efficacy, interventions, preventing

The review specifically addressed patients who had undergone surgery for a fragility hip fracture, a group that remains vulnerable to subsequent fractures, persistent pain and incomplete functional recovery even after the fracture has been stabilized. The authors searched PubMed, Embase, Web of Science and the Cochrane Library for studies published from January 2015 through April 15, 2026. Eligible interventions included zoledronic acid, denosumab, teriparatide, alendronate, risedronate, a sequence of pre-fracture alendronate followed by postoperative teriparatide, and control regimens based on calcium, vitamin D, placebo or a combination of these.

Seven randomized controlled trials and 11 non-randomized comparative studies met the criteria. The researchers used Bayesian random-effects consistency models to combine direct and indirect evidence, expressing dichotomous outcomes as odds ratios with 95% credible intervals and continuous outcomes as mean differences with the same interval type. Primary outcomes included subsequent hip fracture, fractures at any site and bone mineral density reported either in grams per square centimetre or as T-scores. Harris Hip Score and visual analog scale pain scores were secondary outcomes. Analyses restricted to randomized trials or to exactly 12 months of follow-up were conducted where the available networks permitted them.

For subsequent hip fracture within 12 months, five studies with 599 patients contributed data. Both zoledronic acid and denosumab showed trends toward lower risk than background therapy, and zoledronic acid also showed a trend toward lower risk than teriparatide. For fractures at any site within 12 months, six studies involving 725 patients produced favorable numerical trends for zoledronic acid, denosumab and alendronate against control, as well as several comparisons between active treatments. At 24 months, however, four studies involving 1,072 patients showed no clear differences among zoledronic acid, denosumab, teriparatide and control.

Those re-fracture findings came with substantial statistical and methodological uncertainty. Fractures were uncommon, several treatment comparisons contained no events, and some estimates were so extreme that they were sensitive to the priors used in the model. The credible intervals were correspondingly wide, and alternative prior specifications altered the magnitude of some results. Node-splitting analysis did not detect inconsistency between direct and indirect evidence for the 24-month network, and leave-one-out analyses produced broadly stable findings. Even so, the authors described the risk-reduction results as trends rather than definitive evidence of superiority.

Bone mineral density was likewise characterized by outcome-specific patterns rather than a stable medication advantage. Four randomized studies contributing 301 patients supplied areal density data in grams per square centimetre, while four studies with 368 patients supplied T-score data. Zoledronic acid and teriparatide had relatively favorable numerical rankings in the areal-density analysis, whereas alendronate and teriparatide ranked more favorably by T-score. Neither network showed clear between-treatment differences. The apparent divergence also reflected differences in baseline density, anatomical site, laterality, measurement devices and follow-up periods, which could not be resolved with reliable site-specific analyses.

No clear differences emerged for hip function or pain. Eight studies with 684 patients contributed Harris Hip Score data, and eight studies with 1,074 patients contributed visual analog scale data. The Harris Hip Score analysis showed very high heterogeneity, while heterogeneity for pain scores was also very high. Alendronate had a favorable numerical pattern for Harris Hip Score, and zoledronic acid and denosumab showed favorable patterns for pain, but no comparisons clearly distinguished the treatments. Randomized-only and exact-12-month sensitivity analyses were generally consistent with the primary models. The authors cautioned that these outcomes capture the combined effects of surgery, rehabilitation, analgesia, fracture characteristics and baseline health rather than the isolated action of an osteoporosis drug.

Risk-of-bias assessments reinforced the need for caution. All 11 non-randomized studies were judged to have a high overall risk of bias, with concerns involving participant selection, intervention classification, departures from intended treatment, outcome measurement and selective reporting. Although the non-randomized studies received moderate-to-good scores on a supplementary methodological-quality scale, the randomized trials also had unclear or high risk in some areas of sequence generation, allocation concealment and blinding. The GRADE assessment indicated limited certainty across the outcomes, while publication bias could not be assessed reliably because the networks were small and comparison-adjusted funnel plots were not used.

Treatment timing was another unresolved influence. The interval between fracture or surgery and the start of medication could affect early skeletal recovery and the time available for bone-density gains, particularly when antiresorptive and anabolic therapies are compared. Reporting was too inconsistent to form comparable groups, and the evidence could not support reliable subgroup or network meta-regression analyses. The synthesis also did not systematically evaluate adverse events, treatment discontinuation, adherence, renal considerations or hypocalcemia. As a result, it cannot establish comparative benefit-risk profiles, and the relatively favorable signals for individual drugs should not be read as prescribing advice or as evidence that one treatment should be selected over another.

The central message is that the available evidence does not support a universal ranking of osteoporosis medications after fragility hip fracture surgery. Favorable early re-fracture signals for zoledronic acid, denosumab and alendronate were outcome- and time-specific, whereas teriparatide’s relative pattern for bone-mass restoration emerged over longer follow-up. None of these signals was accompanied by consistent advantages across fracture, density, function or pain outcomes. More rigorous head-to-head randomized trials, standardized outcome measurement, clearly reported treatment timing and systematic assessment of safety and adherence are needed before comparative claims can become more reliable.

Subject of Research: Comparative effects of anti-osteoporosis medications after surgery for fragility hip fracture

Article Title: Effects of different anti-osteoporosis medications on re-fracture risk, bone mineral density, and functional recovery following hip fragility fracture surgery: a systematic review and network meta-analysis

Article References: Tian, J., Zhang, H., Lu, Y., Wu, H., Lu, Y., Yu, D., Yu, L., & Luo, Z. (2026). Effects of different anti-osteoporosis medications on re-fracture risk, bone mineral density, and functional recovery following hip fragility fracture surgery: a systematic review and network meta-analysis. Frontiers in Endocrinology, 17, Article 1949781. https://doi.org/10.3389/fendo.2026.1949781

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1949781

Keywords: fragility hip fracture, osteoporosis, network meta-analysis, anti-osteoporosis drugs, re-fracture prevention, bone mineral density, hip function, postoperative pain, evidence certainty, Effects, different, anti-osteoporosis

Electroacupuncture plus an SSRI was also associated with better short-term symptom reduction than SSRIs alone, with a mean difference of −2.98 HAMD points and a 95% credible interval of −3.66 to −2.31. Manual acupuncture combined with serotonin–norepinephrine reuptake inhibitors, or SNRIs, also favored combination treatment over SNRI monotherapy, but the estimate was less precise at −3.16 points, with a wide 95% credible interval of −6.03 to −0.28. These were network estimates combining direct and indirect evidence rather than results from a single head-to-head trial.

The manual-acupuncture-plus-SSRI treatment had the highest posterior probability of occupying the most effective position in the network, Rank 8, at 0.466. Manual acupuncture plus SNRIs followed at 0.406, while electroacupuncture plus SSRIs had a Rank 8 probability of 0.076. SSRI monotherapy had a probability of 0.000 for that position. These figures describe the probability that a treatment would occupy a particular rank after uncertainty was incorporated; they are not summary scores of treatment benefit, and a favorable rank does not by itself demonstrate clinical superiority.

The apparent advantage of manual acupuncture did not survive unchanged under a key sensitivity analysis. Because the trials used both the 17-item and 24-item versions of the HAMD, whose score ranges differ, the researchers repeated the analysis using standardized mean differences. In that model, electroacupuncture plus SSRIs had a standardized mean difference of −0.764, with a 95% credible interval of −1.030 to −0.498, relative to SSRI monotherapy. Manual acupuncture plus SSRIs had a standardized mean difference of −0.659, with a 95% credible interval of −0.842 to −0.476. Both combinations therefore remained associated with improvement over SSRI monotherapy, but their relative order reversed.

The underlying literature was broad but clinically uneven. All included studies were conducted in China and had been published between 2003 and 2025. Participants had primary depression or depression associated with post-stroke, postpartum, perimenopausal, cancer-related, chronic-pain and other clinical settings. Treatment lasted from two to 12 weeks, most commonly four to eight weeks. Acupuncture practices also varied: trials used manual or electroacupuncture and, in some cases, other needling or warming methods, with differing acupoints, treatment frequencies, session durations and antidepressant doses. The researchers therefore considered the transitivity assumption connecting the network comparisons to be uncertain and did not perform subgroup or meta-regression analyses by clinical context.

Quality concerns limit confidence in the treatment estimates. Using the Cochrane Risk of Bias 2 tool, the reviewers found concerns about the randomisation process in 66 studies, commonly because allocation concealment was inadequately reported. Deviations from intended interventions raised concerns in 59 studies, often because the roles and standardisation of acupuncture practitioners were insufficiently described. Outcome measurement raised concerns in 59 trials and was rated at high risk in seven, while three studies were considered at high risk for selective reporting. Although 57 studies were rated as low risk for missing outcome data, the overall body of evidence is vulnerable to biases that could exaggerate or obscure the observed effects.

The statistical model converged satisfactorily, and node-splitting detected no significant local inconsistency for most comparisons that could be tested. However, it identified inconsistency between direct and indirect evidence for electroacupuncture plus SSRIs versus manual acupuncture plus SSRIs, with a reported p value of 0.0063. That particular comparison should therefore be viewed cautiously. The investigators also could not establish transitivity with confidence because potential modifiers—including baseline depression severity, treatment duration, clinical setting and intervention details—were incompletely or unevenly reported. The model’s numerical precision should not be confused with certainty about the clinical importance of the effects.

Safety evidence was too sparse to support a claim that acupuncture-antidepressant combinations were safer, less harmful or equivalent to antidepressant monotherapy. The available networks did not show statistically clear differences in nausea and vomiting, dry mouth or headache for manual acupuncture plus SSRIs or electroacupuncture plus SSRIs relative to SSRIs alone, but credible intervals were extremely wide. Some odds ratios for nausea and vomiting were implausibly small, extending across many orders of magnitude; the authors attributed these estimates mainly to sparse events and zero-event cells rather than interpreting them as evidence of strong protection. Treatment discontinuation was reported too infrequently for a meaningful quantitative synthesis.

The review’s central result is therefore a provisional signal for short-term symptom improvement, not a prescription for a preferred treatment. Manual acupuncture plus SSRIs produced the largest raw-score reduction in the principal network, and both manual and electroacupuncture combinations retained a benefit over SSRI monotherapy in the standardized sensitivity analysis. Yet the shift in their relative ordering, local inconsistency, uncertain transitivity, risk-of-bias concerns, heterogeneous clinical populations and lack of long follow-up prevent a definitive treatment hierarchy. Preclinical studies have proposed effects on serotonergic transmission, stress-axis regulation, neuroimmune signalling and mood-related brain circuits, but such mechanisms remain supportive context and do not prove that acupuncture and antidepressants act synergistically in patients.

Subject of Research: Short-term efficacy and safety of acupuncture–antidepressant combinations for clinically diagnosed depressive disorders in China

Article Title: Acupuncture–antidepressant combinations for clinically diagnosed depressive disorders in China: a systematic review and network meta-analysis

Article References: Qiu, Q., Zhang, Z., Yang, W., Tew, W. Y., Liu, C., Leong Bin Abdullah, M. F. I., Yam, M. F., & Yang, S. (2026). Acupuncture–antidepressant combinations for clinically diagnosed depressive disorders in China: a systematic review and network meta-analysis. Frontiers in Psychiatry, 17, Article 1805745. https://doi.org/10.3389/fpsyt.2026.1805745

Image Credits: AI Generated

DOI: 10.3389/fpsyt.2026.1805745

Keywords: acupuncture, antidepressants, major depressive disorder, network meta-analysis, selective serotonin reuptake inhibitors, electroacupuncture, depression symptoms, systematic review, randomized controlled trials, treatment safety, antidepressant, combinations

The study, led by Yihang Gong of Dankook University in South Korea alongside colleagues from Harbin Sport University, Shenzhen Hospital of Guangzhou University of Chinese Medicine, and Dankook University, searched six databases from inception to July 2026. The team evaluated four primary endpoints: estimated glomerular filtration rate (eGFR), a key measure of kidney filtering capacity; urinary albumin excretion (UAE), a marker of kidney damage; glycated hemoglobin (HbA1c), reflecting long-term blood sugar control; and fasting blood glucose (FBG). By mapping these interventions into a network meta-analysis, the researchers could compare modalities both directly, where head-to-head trials existed, and indirectly through shared comparators.

The headline finding concerns resistance training. Compared with usual care, resistance training was the sole modality that significantly increased eGFR, with a mean difference of 11.36 mL/min/1.73 m², and it also significantly outperformed aerobic training head-to-head by 8.61 mL/min/1.73 m². In a disease where declining filtration rate signals progressive kidney failure, an improvement of that magnitude would be remarkable, potentially rivaling pharmacological interventions. On paper, resistance training also ranked highest across nearly every outcome, achieving a SUCRA score of 98.8 percent for eGFR, 94.2 percent for HbA1c, and 86.1 percent for fasting glucose.

But the authors urge heavy caution before rewriting clinical guidelines around those numbers. When the two trials judged to be at high risk of bias were excluded in sensitivity analyses, the eGFR benefit of resistance training lost statistical significance entirely, shrinking to a non-significant 8.24 mL/min/1.73 m². The researchers also note that an effect of this size is disproportionately large relative to established nephroprotective drugs, and that in early diabetic nephropathy, which made up 60 percent of the included trials, acute filtration increases may partly reflect exercise-induced glomerular hyperfiltration or transient hemodynamic shifts rather than genuine structural kidney preservation. Disparate filtration equations across trials, from MDRD to CKD-EPI, add further measurement noise.

Mind-body exercise carved out its own niche. It was the only intervention that significantly reduced urinary albumin excretion compared with usual care, with a standardized mean difference of -1.41. Albuminuria is one of the earliest and most prognostically important signs of diabetic kidney damage, so a practice like Tai Chi or Baduanjin, low-impact and highly accessible, offering a measurable reduction is clinically intriguing. The certainty of this evidence was rated low under the CINeMA framework, but the finding aligns with the supplement’s exploratory observation that mind-body exercise also produced a notable systolic blood pressure reduction of roughly 15.7 mmHg, reinforcing its potential cardiovascular relevance.

For glycemic control, the picture was broader. Resistance training, aerobic training, and combined training all significantly lowered HbA1c compared with usual care, by 1.65 percent, 1.07 percent, and 0.70 percent respectively, though all at low certainty. Fasting blood glucose fell significantly only with resistance training (-1.78 mmol/L) and combined training (-1.71 mmol/L), the latter supported by moderate certainty of evidence, the highest rating in the entire analysis. In head-to-head comparisons, both resistance and combined training significantly beat aerobic training on fasting glucose.

Combined training, which blends aerobic and resistance elements, emerged as the most durable performer in subgroup analyses. It was the only modality to sustain significant fasting glucose reductions across both short-term regimens of twelve weeks or less (-1.50 mmol/L) and longer-term programs beyond twelve weeks (-2.39 mmol/L). Most other benefits, including eGFR and albumin improvements from resistance and aerobic training, faded in the long-term subgroup. This temporal pattern matters because diabetic kidney disease is a chronic, progressive condition, and interventions whose effects evaporate after a few months offer limited clinical value.

Sensitivity analyses probing confounding added further nuance. When the researchers restricted the synthesis to pure exercise interventions, excluding four trials that bundled exercise with traditional Chinese medicine, specialized nursing protocols, or structured nutritional guidance, the HbA1c benefit of combined training lost significance. Yet in a pragmatic scenario that removed medication and nursing trials while retaining guideline-directed dietary advice, all primary conclusions held. The authors interpret this as evidence that routine nutritional counseling, an intrinsic part of standard diabetic kidney disease care, does not undermine the exercise signal, though residual interactive confounding cannot be fully excluded.

Perhaps the most sobering finding concerns safety. Structured adverse event monitoring was documented in only five of the twenty trials, or 25 percent. Three of those reported mild, transient events such as hypoglycemia, musculoskeletal discomfort, and falls without serious harm, while two explicitly declared zero adverse events. The remaining fifteen trials simply omitted safety surveillance altogether. This sparse reporting means the comparative safety of different exercise intensities and modalities in diabetic kidney disease cannot be definitively established, a gap the authors say mandates routine surveillance in future studies.

The practical takeaway is individualization rather than a single winning prescription. The authors recommend combined training as a balanced, pragmatic foundation for sustained glycemic management, mind-body exercise as a viable strategy for patients with limited functional capacity or concerns about albuminuria, and resistance training only with caution given its fragile renal evidence. Direct head-to-head trials between active modalities remain scarce, with only three in the entire network, so most comparative rankings rest on indirect evidence. Larger, multi-center trials with standardized FITT reporting, longer follow-up, and mandatory safety monitoring are needed before exercise prescriptions for diabetic kidney disease can be built on firmer ground.

Subject of Research: Comparative efficacy of exercise modalities on renal function and glycemic control in non-dialysis diabetic kidney disease

Article Title: Comparative efficacy of different exercise modalities on renal function and glycemic control in patients with non-dialysis diabetic kidney disease: a systematic review and network meta-analysis

Article References: Gong, Y., Zheng, J., Huang, W., & Jeon, Y. (2026). Comparative efficacy of different exercise modalities on renal function and glycemic control in patients with non-dialysis diabetic kidney disease: a systematic review and network meta-analysis. Frontiers in Endocrinology, 17, Article 1991940. https://doi.org/10.3389/fendo.2026.1991940

Image Credits: AI Generated

DOI: 10.3389/fendo.2026.1991940

Keywords: diabetic kidney disease, exercise therapy, resistance training, network meta-analysis, renal function, glycemic control, HbA1c, albuminuria, mind-body exercise, combined training, eGFR, randomized controlled trials

The research team confronted a problem that plagues the entire exercise-on-the-brain literature: dozens of randomized trials have tested everything from aerobic training to Tai Chi, but almost no head-to-head comparisons exist. Standard meta-analyses can only pool trials that compare the same two interventions, leaving clinicians guessing about which modality actually works best. Network meta-analysis, or NMA, solves this by linking interventions across a web of trials, using both direct comparisons and indirect evidence transmitted through common comparators. Su and colleagues performed Bayesian random-effects NMAs, a statistical framework in which prior distributions are combined with trial data to generate posterior estimates and credible intervals, allowing every intervention to be ranked against every other even when they have never been directly compared.

The evidence base was assembled with unusual rigor. Following PRISMA-NMA guidance and with the review prospectively registered in PROSPERO, the team searched PubMed, Embase, the Cochrane Library, Web of Science, MEDLINE via Ovid, Scopus, CNKI, Wanfang, VIP, Google Scholar, and ProQuest from inception to August 16, 2026. They included randomized controlled trials involving older adults with cognitive frailty and assessed risk of bias with the RoB 2 tool, while certainty of evidence was evaluated using CINeMA, a framework designed specifically for grading confidence in network meta-analytic estimates. The final dataset comprised nineteen publications representing seventeen independent randomized trial datasets and 1,691 unique participants.

Five outcomes anchored the analysis, each chosen for its clinical relevance. Cognitive function was measured with the Montreal Cognitive Assessment (MoCA) and the Mini-Mental State Examination (MMSE), while physical function was captured through the Timed Up and Go test (TUG), grip strength, and the Short Physical Performance Battery (SPPB). The interventions compared spanned an impressive range: multicomponent exercise, Baduanjin, Baduanjin combined with cognitive training, motor-cognitive training, exercise-based cognitive combined training, Otago exercise, Tai Chi, resistance training, high-intensity interval training, aerobic exercise, and daily living rehabilitation training.

The headline cognitive result is a study in statistical subtlety. For MoCA, multicomponent exercise showed the most favorable ranking profile, with a mean difference of 4.81 points and a 95 percent credible interval spanning −0.27 to 10.66. Yet it was Baduanjin combined with cognitive training that produced the only estimate whose credible interval excluded the null, with a mean difference of 4.25 points (95 percent credible interval 0.22 to 8.39). In plain terms, the ranking crowned multicomponent exercise as the leading node, but only the Baduanjin-plus-cognitive-training combination delivered an estimate precise enough to rule out chance. For context, a MoCA mean difference above four points is substantial in a population where total scores range from zero to thirty, so both signals deserve attention from trial designers.

For the MMSE, daily living rehabilitation training ranked highest; for the Timed Up and Go test, Otago exercise led the field; for grip strength, multicomponent exercise again topped the rankings; and for the SPPB, Tai Chi came out on top. But here the story turns cautionary: the credible intervals for all of these leading estimates included the null. The authors are explicit that these rankings should be read as potentially favorable, outcome-specific signals rather than definitive prescriptions. In a Bayesian network meta-analysis, a credible interval that crosses zero means the data cannot exclude the possibility that the intervention’s true effect is nothing at all, no matter how high its rank.

The technical limitations of the evidence network are as informative as the results themselves. Substantial heterogeneity was observed for MoCA, meaning the trials feeding that network varied widely in populations, intervention dose, and study design, inflating uncertainty in the pooled estimates. Several physical-function networks were sparse, with few trials connecting the intervention nodes, and formal inconsistency assessment, the NMA equivalent of checking whether direct and indirect evidence agree, was frequently not possible because the network geometry lacked closed loops. Sensitivity analyses generally preserved the main ranking patterns, a reassuring sign, although several leading nodes rested on limited evidence.

Certainty of evidence, graded with CINeMA, was predominantly moderate to low, with some comparisons rated as very low. This grading reflects the interplay of imprecision, heterogeneity, and network sparsity rather than any single flaw. The authors emphasize that the findings should inform future comparative trials rather than be interpreted as definitive evidence for individualized exercise prescriptions. That framing matters: network meta-analysis is a hypothesis-generating tool, and the ranking of multicomponent exercise and Baduanjin combined with cognitive training at the top of cognitive networks is best seen as a signal about where the next generation of adequately powered head-to-head trials should concentrate.

Why does this matter for the broader science of aging? Cognitive frailty sits at the intersection of two degenerative trajectories, and the coexistence of physical and cognitive decline suggests shared biological mechanisms, from chronic inflammation and vascular dysfunction to sarcopenia-related signaling. Exercise interventions that simultaneously challenge the cardiovascular system, the musculoskeletal system, and cognition, as multicomponent programs and mind-body practices like Baduanjin and Tai Chi do, may engage these mechanisms in parallel. The fact that mind-body exercise combined with explicit cognitive training produced the only statistically robust cognitive signal hints that pairing physical movement with deliberate cognitive challenge could be more effective than either component alone, a hypothesis that motor-cognitive training research has been circling for years.

The work, supported by the Shanghai Nursing Association Scientific Research Project and the Pudong New Area Health Commission, arrives at a moment when populations worldwide are aging rapidly and health systems are searching for scalable, non-pharmacological interventions. Exercise is cheap, safe, and adaptable, and if the right modality can be identified for cognitive frailty, the public health payoff could be enormous. Su and colleagues have not handed clinicians a prescription, but they have produced something arguably more valuable: a rigorous, transparent map of what the evidence does and does not show, and a clear agenda for the trials that must come next. For the 1,691 participants whose data underpin this analysis and the millions of older adults they represent, that map may prove to be the first step on a path toward genuinely reversible cognitive decline.

Subject of Research: Comparative effects of exercise-based interventions on cognitive and physical function in older adults with cognitive frailty

Article Title: Exercise-based interventions for cognitive frailty in older adults: a systematic review and network meta-analysis

Article References: Su, L., Abudoureheman, W., Jin, Y., Wang, Y., Bai, S., & Liu, S. (2026). Exercise-based interventions for cognitive frailty in older adults: a systematic review and network meta-analysis. BMC Geriatrics. https://doi.org/10.1186/s12877-026-08435-1

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08435-1

Keywords: cognitive frailty, exercise intervention, network meta-analysis, older adults, cognitive function, physical function, BMC Geriatrics, Exercise-based, interventions, cognitive, frailty, older

Despite this biological logic, the clinical evidence base for interventions that target mitochondria in older people with type 2 diabetes remains scattered and difficult to interpret. Hundreds of trials have tested drugs, supplements, and lifestyle programs that plausibly act on mitochondrial pathways, but few studies have focused specifically on adults aged 60 and above, and almost none have compared these interventions against one another in a rigorous, quantitative way. A research team led by Hao Liu, Aochuan Sun, Yu Deng, Yihan Li, Jiayu Hu, and Zhengtang Liu now aims to change that. Their protocol, published in PLOS One, lays out a systematic review with a conditional network meta-analysis designed to map and rank the comparative efficacy and safety of mitochondrial-targeting interventions in this underexplored population.

The scope of the planned review is deliberately broad. The researchers define interventions with mitochondrial effects in two ways: either the intervention’s primary mechanism of action is known to target mitochondrial pathways, or the intervention has been explicitly evaluated in trials for its effects on direct markers of mitochondrial function. This umbrella captures pharmacological agents, nutritional supplements, and lifestyle modifications, all of which will be compared against placebo, standard care, or no intervention at all. By casting a wide net, the team hopes to assemble a comprehensive picture of how different therapeutic strategies, from common medications to exercise regimens, influence the cellular machinery that governs energy metabolism in aging bodies.

The primary endpoint of the review is mitochondrial function itself, measured through biomarkers that will be categorized by biological domain, tissue source, and assay method. This classification matters because mitochondrial health can be assessed in strikingly different ways: some studies measure respiratory chain activity in muscle biopsies, others track circulating markers of mitochondrial damage, and still others rely on indirect proxies. By organizing outcomes according to where and how they were measured, the researchers aim to bring analytical order to a heterogeneous literature. Secondary outcomes extend into clinically familiar territory, including glycated hemoglobin, fasting plasma glucose, fasting insulin, the homeostatic model assessment of insulin resistance, body weight, and body mass index, alongside broader measures of geriatric health and safety.

Methodologically, the protocol follows the highest current standards for evidence synthesis. It was prepared in accordance with the PRISMA-P reporting guideline for protocols and will adhere to PRISMA 2020 and, where applicable, the PRISMA-NMA extension for network meta-analyses. The review is registered with PROSPERO under identifier CRD420251145356, a step that guards against selective reporting and undisclosed changes of plan. The search strategy spans five major databases, PubMed, Embase, the Cochrane Central Register of Controlled Trials, Web of Science, and Scopus, covering the literature from inception to April 2026. Search terms focus on type 2 diabetes, mitochondrial function and dysfunction, and interventions, with age eligibility assessed during full-text screening.

Age criteria receive careful treatment because many diabetes trials enroll broad adult populations without reporting age-specific results. The primary analysis will include trials whose participants are aged 60 or older, or trials from which data for this age group can be extracted. As a sensitivity analysis, the team will also examine studies with a mean or median participant age of at least 60, allowing them to test whether conclusions hold under a slightly looser definition of the target population. Throughout the review, two independent reviewers will handle study screening, data extraction, and risk-of-bias assessment using the Cochrane RoB 2 tool, a structured instrument for evaluating the reliability of randomized trial results.

The analytical centerpiece, the network meta-analysis, represents a statistical advance over traditional pairwise comparisons. Conventional meta-analyses can only pool trials that compare the same two interventions directly, leaving many treatments unconnected. A network meta-analysis instead weaves together direct and indirect evidence across an entire network of trials, enabling simultaneous comparison and ranking of many interventions even when some have never been tested head to head. Crucially, the team has built a conditional into their design: the network meta-analysis will proceed only if the evidence network is sufficiently connected and clinically coherent. If those conditions are not met, the researchers will fall back on pairwise meta-analysis or a structured narrative synthesis, ensuring that statistical ambition never outruns the underlying data.

Heterogeneity, the variability among trial results that can arise from differences in populations, interventions, and measurement methods, poses a well-known threat to any pooled analysis. The protocol addresses it on several fronts. Random-effects models will be employed throughout, acknowledging that true effects may vary across studies. Prespecified subgroup analyses and meta-regression will explore potential sources of heterogeneity, such as intervention duration and baseline glycated hemoglobin levels. If significant heterogeneity persists even after these efforts, the team will apply the Synthesis Without Meta-analysis framework, a formal approach for narrative review that avoids forcing incompatible results into a single statistical estimate. Finally, the certainty of the evidence will be appraised using the GRADE system and the CINeMA tool, both of which grade how much confidence readers can place in each conclusion.

The ethical footprint of the project is minimal by design, since the review will rely exclusively on previously anonymized data from primary studies, and the findings are intended for publication in a peer-reviewed journal. Yet the potential implications extend well beyond academic bookkeeping. If the network meta-analysis succeeds in ranking mitochondrial interventions by efficacy and safety, clinicians caring for older adults with type 2 diabetes could gain an evidence-based hierarchy of options, from pharmacological agents to nutritional supplements to lifestyle changes, each anchored to measurable effects on cellular energy metabolism. For a population in which diabetes complications, frailty, and polypharmacy intersect, such clarity could translate into more precise and personalized treatment decisions.

The study also highlights a broader shift in diabetes research: the recognition that aging and metabolism are intertwined at the subcellular level. As the global population of older adults with type 2 diabetes continues to grow, interventions that restore mitochondrial function may offer benefits that conventional glucose-lowering drugs alone cannot capture, potentially affecting fatigue, physical function, and the trajectory of diabetic complications. Whether any specific intervention will emerge as a clear frontrunner remains to be seen, and the conditional design of the planned analysis reflects an appropriate scientific caution. What is already clear is that the era of treating diabetes without reference to the cell’s power stations is drawing to a close, and rigorous, age-focused evidence synthesis is the tool that will determine which mitochondrial strategies deserve a place in clinical practice.

Subject of Research: Systematic review and network meta-analysis protocol evaluating mitochondrial-targeting interventions for older adults with type 2 diabetes

Article Title: Interventions with mitochondrial-related effects in older adults with type 2 diabetes: Protocol for a systematic review with conditional network meta-analysis

Article References: Liu, H., Sun, A., Deng, Y., Li, Y., Hu, J., & Liu, Z. (2026). Interventions with mitochondrial-related effects in older adults with type 2 diabetes: Protocol for a systematic review with conditional network meta-analysis. PLOS One, 21(10), e0360683. https://doi.org/10.1371/journal.pone.0360683

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0360683

Keywords: mitochondria, type 2 diabetes, older adults, network meta-analysis, systematic review, insulin resistance, mitochondrial dysfunction, glycemic control, PRISMA, PROSPERO, evidence synthesis, biomarkers

The research team, led by investigators at Beijing University of Chinese Medicine and Peking University Third Hospital, systematically searched PubMed, Embase, and the Cochrane Library through 6 November 2025. From an initial haul of 1,180 records, they identified ten randomized controlled trials encompassing 3,959 patients with metastatic colorectal cancer who had progressed after second-line therapy. Each trial compared one of the three oral agents—alone or, in some cases, TAS-102 combined with bevacizumab—against placebo or best supportive care. Risk of bias was assessed with the Cochrane RoB-2 tool, and the protocol was prospectively registered with PROSPERO.

Because no trial has ever randomized patients directly between regorafenib, fruquintinib, and TAS-102, the researchers turned to network meta-analysis, a statistical framework that uses a common comparator—in this case placebo—to make indirect comparisons across trials. They fitted a Bayesian random-effects consistency model with non-informative normal priors, ran four parallel Markov chains through 20,000 adaptation iterations and 50,000 sampling iterations, and confirmed convergence using the Gelman-Rubin diagnostic. The resulting network was star-shaped, anchored on placebo, with no closed loops among the active treatments, which meant formal inconsistency testing was impossible; instead, the team relied on transitivity checks comparing trial characteristics and potential effect modifiers.

The headline finding for the overall population was straightforward: every active treatment significantly prolonged overall survival compared with placebo. Fruquintinib produced a hazard ratio of 0.66 (95% credible interval 0.53–0.84), regorafenib 0.69 (0.51–0.88), TAS-102 0.69 (0.55–0.85), and TAS-102 plus bevacizumab 0.41 (0.28–0.59). The combination of TAS-102 with bevacizumab ranked first by SUCRA—the surface under the cumulative ranking curve—while the three single agents showed broadly similar efficacy, with indirect differences among them not consistently statistically significant. Progression-free survival followed the same pattern, with all experimental arms outperforming placebo and the doublet again delivering the largest benefit.

One indirect comparison did stand out: fruquintinib showed a more favorable progression-free survival than TAS-102, with a hazard ratio of 0.65 (95% CrI 0.45–0.96), placing it highest among the monotherapies. The authors caution, however, that this estimate rests entirely on indirect evidence in the absence of head-to-head trials and should be interpreted cautiously. Heterogeneity also varied considerably across the network: it was essentially absent for fruquintinib versus placebo (I² ≈ 0%), moderate for TAS-102 (OS I² = 21.8%), and high for regorafenib (OS I² = 67.5%, PFS I² = 83.1%), prompting the use of random-effects models throughout.

The most novel contribution of the analysis lies in its KRAS stratification. During transitivity checks, KRAS mutational status emerged as the most plausible source of the between-study heterogeneity, driven largely by differences in baseline KRAS prevalence—Asian trials such as CONCUR and TERRA enrolled 34% and 37% KRAS-mutant patients respectively, while the remaining trials ranged from 40% to 70%. When the team restricted the analysis to trials reporting KRAS-subgroup overall survival data, heterogeneity collapsed dramatically in the wild-type subgroup, with I² falling to 20% or below for every comparison, and the three agents ranked almost side by side with no meaningful differences between them.

In the KRAS-mutant subgroup, the picture shifted subtly. Fruquintinib and TAS-102 each produced numerically favorable overall survival estimates versus placebo—hazard ratios of 0.71 (95% CrI 0.51–1.01) and 0.74 (95% CrI 0.50–0.99) respectively—whereas regorafenib did not reach statistical significance, with a hazard ratio of 0.81 (95% CrI 0.52–1.19). The exploratory ranking placed fruquintinib first, TAS-102 second, and regorafenib last. The authors are emphatic, however, that these differences in statistical significance versus placebo do not demonstrate superiority of fruquintinib or TAS-102 over regorafenib, and that the rank probabilities are hypothesis-generating rather than evidence of a predictive effect of KRAS status.

The mechanistic backdrop helps explain why KRAS status might matter. KRAS is mutated in approximately 40% of colorectal cancers, driving constitutive RAS signaling and unrestrained cell growth. Prior laboratory work has suggested that exon-2 KRAS mutations, particularly G12C and G12A, maintain elevated BRAF-ERK-ELK phosphorylation and blunt regorafenib-induced apoptosis and vessel shutdown, while a recently described KRAS/ERK2–TRIM21–c-Myc/ENO2 glycolytic axis promotes regorafenib resistance. Notably, among the three TAS-102 trials, the Japanese phase 2 study by Yoshino and colleagues showed a markedly stronger benefit in KRAS-mutant patients (hazard ratio 0.44) than in wild-type disease (0.70), a pattern that largely accounts for the residual heterogeneity in that treatment node.

The combination of TAS-102 with bevacizumab emerged as the standout regimen overall. Mechanistically, the pairing is plausible: trifluridine inhibits thymidylate synthase and, through its metabolite FTY, becomes incorporated into DNA causing miscoding, while tipiracil protects trifluridine from degradation by blocking thymidine phosphorylase; bevacizumab, meanwhile, suppresses VEGF-driven angiogenesis and remodels the tumor microenvironment. The authors suggest these mechanisms may act synergistically, and a separate analysis indicated the combination conferred a particularly pronounced survival benefit in RAS-mutant patients (hazard ratio 0.38, 95% CrI 0.19–0.79) with no significant advantage over monotherapy in the wild-type population—though only one study contributed to that subgroup node, precluding a stable network estimate.

The study has clear limitations that the authors acknowledge candidly. No head-to-head randomized comparisons exist among the three agents, the star-shaped network precluded formal inconsistency testing, and individual-participant data were unavailable, so a formal treatment-by-KRAS interaction test could not be performed. Comparative safety was not formally meta-analyzed because safety definitions and reporting thresholds differed substantially across trials, and progression-free survival subgroup analysis was abandoned due to insufficient sample size and high heterogeneity (I² ranging from 77.7% to 92.2%). Publication-bias testing was deemed unreliable with fewer than ten trials per outcome. The authors conclude that for KRAS-wild-type patients, standard third-line regimens or TAS-102 plus bevacizumab are reasonable, while for KRAS-mutant patients fruquintinib may represent a sensible later-line option—pending confirmation from prospective, KRAS-stratified or head-to-head trials that the field has yet to deliver.

Subject of Research: Comparative efficacy of third-line oral therapies for KRAS-mutant metastatic colorectal cancer

Article Title: Comparative efficacy of fruquintinib, regorafenib or TAS-102 in patients with KRAS-mutant metastatic colorectal cancer: a systematic review and network meta-analysis

Article References: Zhou, Y., Zhao, S., Zheng, X., Wang, Y., Wang, M., Chen, S., Wu, Z., Zang, D., & Si, W. (2026). Comparative efficacy of fruquintinib, regorafenib or TAS-102 in patients with KRAS-mutant metastatic colorectal cancer: a systematic review and network meta-analysis. Frontiers in Oncology, 16, Article 1948508. https://doi.org/10.3389/fonc.2026.1948508

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1948508

Keywords: fruquintinib, regorafenib, TAS-102, KRAS mutation, metastatic colorectal cancer, network meta-analysis, overall survival, progression-free survival, bevacizumab, third-line therapy, precision oncology, Bayesian analysis

Acute severe ulcerative colitis is a potentially life-threatening emergency marked by frequent bloody stools, systemic toxicity, and the need for urgent hospitalization. It develops in up to one-quarter of ulcerative colitis patients during their disease course. Intravenous corticosteroids remain the recommended first line of attack, yet roughly 30 to 40 percent of patients fail to respond adequately after three to five days. At that point, clinicians must choose quickly between medical rescue therapy and colectomy, because delayed surgery or ineffective drug treatment can compound the danger.

For decades, the two established rescue options have been infliximab, a monoclonal antibody that neutralizes tumor necrosis factor, and calcineurin inhibitors such as ciclosporin. Landmark randomized trials, including the CySIF trial and the pragmatic CONSTRUCT study, found no meaningful difference between the two drugs in treatment failure, colectomy, or serious adverse events. But the therapeutic landscape has grown more complicated. Accelerated and high-dose infliximab regimens have been proposed to counteract the rapid drug clearance seen in severely inflamed patients, while oral Janus kinase inhibitors such as tofacitinib and upadacitinib have emerged as candidates, particularly for patients already exposed to biologic therapy.

To untangle this crowded field, an international team of researchers led by investigators at Brookdale University Hospital and Medical Center in New York, with collaborators spanning institutions in the United States, Egypt, and beyond, searched five major electronic databases from inception through June 30, 2026. They screened 5,498 records and ultimately included 18 studies: three randomized controlled trials and 15 comparative observational studies. The team conducted separate frequentist random-effects network meta-analyses for four outcomes: colectomy at approximately 12 to 14 weeks, colectomy during hospitalization or within 30 days, colectomy at approximately 12 months, and serious adverse events. The review was registered with PROSPERO under identifier CRD420261412884.

The headline result concerned the classic head-to-head comparison. At the short-term follow-up window of roughly 12 to 14 weeks, ciclosporin was associated with significantly higher odds of colectomy than standard-dose infliximab, with an odds ratio of 2.19 and a 95 percent confidence interval of 1.17 to 4.12. The pattern held at earlier and later time points: ciclosporin showed roughly three-fold higher odds of early colectomy, defined as surgery during the index hospitalization or within 30 days, and more than three-fold higher odds of colectomy at 12 months. Notably, serious adverse events were statistically indistinguishable between the two drugs, with an odds ratio of 0.94.

By contrast, neither accelerated infliximab induction, in which the second and third infusions are given earlier than the conventional week-2 and week-6 schedule, nor high-dose infliximab at 10 milligrams per kilogram showed a significant advantage over the standard 5-milligram-per-kilogram regimen. The confidence intervals for these intensified strategies were strikingly wide, reflecting small treatment groups and sparse event counts. Tacrolimus, another calcineurin inhibitor, also did not differ significantly from standard-dose infliximab at either early or 12-month follow-up, though its network estimate rested on a single retrospective study of just 61 patients.

The Janus kinase inhibitor evidence told an intriguing but incomplete story. In a retrospective matched case-control study, tofacitinib was associated with a lower adjusted 90-day colectomy risk than matched standard-care controls among biologic-experienced hospitalized patients, with an adjusted hazard ratio of 0.28. A separate international comparison of tofacitinib and upadacitinib found nearly identical colectomy-free survival, 79 percent versus 80 percent at day 98, with no difference between the survival curves. However, these studies could not be validly connected to the main infliximab-ciclosporin network because they lacked a common comparator, so no global ranking across drug classes was possible.

Beneath the surface of the pooled estimates lies a critical tension. The apparent efficacy advantage of infliximab over ciclosporin was driven predominantly by observational studies, which pooled to an odds ratio of 2.88 favoring infliximab, whereas the direct randomized comparison in the Laharie trial was essentially neutral, with an odds ratio of 0.78. The authors identify several plausible explanations for this divergence. Many observational cohorts compared patients treated in nonoverlapping calendar periods, meaning ciclosporin cohorts often predated infliximab cohorts and may have benefited less from improvements in supportive care, infection screening, thromboprophylaxis, and surgical decision-making. Moreover, ciclosporin was frequently used as a temporary bridge to thiopurines, while infliximab could be continued as long-term maintenance therapy, potentially skewing 12-month colectomy rates in favor of the latter strategy.

Risk of bias assessments reinforced the need for caution. All three randomized trials were judged to have some concerns, largely because open-label designs could have influenced decisions about additional rescue treatment or the timing of surgery. Among the 15 observational studies, 14 were rated as having serious risk of bias, primarily due to confounding by indication: patients with more severe inflammation, lower albumin, or prior biologic failure were more likely to receive intensified infliximab, tacrolimus, or JAK inhibitors. Using the GRADE framework, the certainty of evidence ranged from low to very low, and no comparison achieved moderate or high certainty.

The practical message for clinicians is nuanced rather than revolutionary. Standard-dose infliximab emerges as an effective and practical default rescue strategy, with the added advantage that it can be continued for maintenance after successful induction. Yet ciclosporin should not be dismissed as obsolete; randomized trials continue to support it as a legitimate option, particularly in thiopurine-naïve patients, those with contraindications to anti-TNF therapy, or centers experienced in therapeutic drug monitoring. The PREDICT-UC randomized trial, which found no benefit of an initial 10-milligram-per-kilogram dose or accelerated induction over standard dosing, argues against routine intensification for all patients, though individualized escalation may still help those with profound hypoalbuminemia or suspected rapid drug clearance.

The authors conclude that adequately powered randomized comparisons are urgently needed, incorporating contemporary biologic-exposed populations and newer small-molecule therapies, with standardized definitions of steroid failure and consistent reporting of colectomy across time points. Until such trials arrive, treatment selection should remain individualized, weighing previous therapy, disease severity, serum albumin, comorbidities, and institutional expertise. And one principle remains nonnegotiable: early surgical consultation should run in parallel with medical rescue planning, because no drug ranking, however sophisticated the statistics, should justify delaying colectomy in a deteriorating patient.

Subject of Research: Comparative efficacy and safety of medical rescue therapies for steroid-refractory acute severe ulcerative colitis

Article Title: Comparative outcomes of advanced medical rescue therapies in steroid-refractory acute severe ulcerative colitis: a systematic review and network meta-analysis

Article References: Alabdallat, M. Y., Al Ta’ani, O., Bakain, T., Al-Dwairy, A., Ghannam, M., Atasi, A., Mashal, F., Moghib, K., Bakheet, N., Jagroo, J. G., Hehr, R., Salem, A., & Cheung, D. (2026). Comparative outcomes of advanced medical rescue therapies in steroid-refractory acute severe ulcerative colitis: a systematic review and network meta-analysis. Frontiers in Gastroenterology, 5, Article 1937925. https://doi.org/10.3389/fgstr.2026.1937925

Image Credits: AI Generated

DOI: 10.3389/fgstr.2026.1937925

Keywords: ulcerative colitis, infliximab, ciclosporin, colectomy, network meta-analysis, rescue therapy, tacrolimus, JAK inhibitors, steroid refractory, acute severe ulcerative colitis, gastroenterology, systematic review

The technical machinery behind the analysis is a frequentist random-effects network meta-analysis, a statistical framework that allows interventions to be compared both directly, within the same trials, and indirectly, across the wider network of studies. Rather than asking whether exercise works in general, the researchers asked which modality works best for which physical function deficit. Rankings were determined using the surface under the cumulative ranking curve, or SUCRA, a metric that assigns each intervention a probability of being the most effective treatment, expressed as a percentage from 0 to 100. The certainty of the evidence supporting each ranking was then evaluated using the GRADE minimally contextualized framework, which grades confidence from high to low based on study limitations, inconsistency, indirectness, imprecision, and publication bias.

The headline finding concerns elastic band resistance training, a low-cost, accessible modality that emerged as the most effective intervention for handgrip strength, one of the most clinically meaningful markers of whole-body muscle function in older adults. Elastic band training increased handgrip strength by a mean difference of 1.95 kilograms, with a 95 percent confidence interval of 1.13 to 2.76 kilograms, and achieved a SUCRA ranking of 98.6 percent, meaning it was almost certainly the top-performing modality for this outcome. Critically, this evidence was rated as moderate certainty under network GRADE, the highest confidence level awarded to any single modality-specific strength finding in the analysis.

Machine-based resistance training showed a different profile. For knee extension strength, a direct measure of quadriceps capacity that underpins standing, stair climbing, and chair rise, machine-based training produced a mean difference of 2.96 kilograms, with a 95 percent confidence interval of 1.59 to 4.34 kilograms and a SUCRA ranking of 65.3 percent, indicating possible intermediate effectiveness. The certainty of this evidence was low, however, meaning clinicians should interpret the magnitude with caution. The distinction matters because it suggests that different resistance modalities may target different strength deficits with different efficiency, a finding that runs counter to the one-size-fits-all approach embedded in many current guidelines.

The dose-response analyses add a layer of prescription precision that few prior syntheses have attempted. For elastic band training, greater strength gains were observed with higher training volume, meaning more total repetitions or sets per session produced progressively larger improvements. For machine-based resistance training, the opposite prescription logic applied: the modality was more effective at high intensity, meaning heavier loads relative to maximum capacity drove the best results. These findings imply that the optimal dose parameters are modality-specific, and that simply telling an older adult to do resistance exercise twice a week misses the point of how the exercise should be configured.

For the broader picture of functional mobility, the ability to walk, transfer, and navigate daily environments, multicomponent training that includes resistance training provided the most comprehensive benefits. Multicomponent programs typically combine resistance, balance, aerobic, and flexibility elements into a single intervention, and the network meta-analysis found that those incorporating resistance work ranked highest for overall functional mobility gains. The certainty of this evidence was low, however, reflecting heterogeneity across the included trials and imprecision in the effect estimates. Still, the finding aligns with the biological rationale that mobility is a compound capacity requiring multiple physiological systems, including muscle strength, power, balance control, and cardiovascular endurance, to be trained simultaneously.

Sarcopenia, the age-related loss of muscle mass and function, and frailty, a state of decreased physiological reserve and heightened vulnerability to stressors, affect a substantial and growing share of the global older population. Exercise remains the primary non-pharmacological intervention for both conditions, yet the comparative effects and optimal prescription parameters of specific modalities have remained unclear despite decades of trials. This analysis addresses that gap by synthesizing evidence from PubMed, Ovid MEDLINE, Embase, ClinicalTrials.gov, PEDro, and the Cochrane Library, with studies included up to June 1, 2025. The systematic review was registered in PROSPERO under registration number CRD420251042815, ensuring transparency in the review protocol.

The clinical implications of the findings are straightforward but consequential. Rather than prescribing exercise generically, clinicians working with older adults with sarcopenia and frailty can now adopt a modality-specific, deficit-targeted approach: if the primary deficit is handgrip strength, elastic band resistance training carries the strongest and most certain evidence; if the deficit is knee extension strength, machine-based training at high intensity may be the better choice, though with lower certainty; and if the goal is overall functional mobility, multicomponent training that includes resistance work offers the most comprehensive benefits. The dose-response findings further refine this logic, suggesting that elastic band programs should prioritize volume while machine-based programs should prioritize intensity.

The GRADE certainty ratings embedded throughout the analysis serve as an honest accounting of what the evidence can and cannot support. Moderate certainty for elastic band training and handgrip strength means the true effect is likely close to the estimated 1.95 kilogram improvement, though further research could still refine the magnitude. Low certainty for machine-based training, multicomponent training and mobility, and other modality-specific findings means the true effects could differ substantially from the point estimates, and that future trials may change the rankings. This transparency is a feature, not a flaw, of the analysis, and it gives clinicians a calibrated sense of how much confidence to place in each prescription decision.

The study was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea under grant NRF-2025S1A5A2A03012772. The corresponding author is Hyuntae Park of Dong-A University, and the research team includes investigators from the Department of Health Sciences and the Digital Healthcare Institute at Dong-A University in Busan, along with Ting-Fu Lai of the School of Gerontology and Long-Term Care at Taipei Medical University in Taiwan. As the global population ages and the burden of sarcopenia and frailty grows, the shift from generic exercise advice to modality-specific, deficit-targeted prescriptions represents a meaningful step toward precision geriatric medicine, where the right exercise, at the right dose, for the right deficit, becomes the standard of care rather than the exception.

Subject of Research: Geriatrics

Article Title: Modality-specific, deficit-targeted exercise prescriptions for physical function in sarcopenia and frailty: a network meta-analysis

Article References: Wang, K., Lai, T.-F., Noh, E.-S., Yang, J., Min, C., Bae, S., & Park, H. (2026). Modality-specific, deficit-targeted exercise prescriptions for physical function in sarcopenia and frailty: a network meta-analysis. BMC Geriatrics. https://doi.org/10.1186/s12877-026-08352-3

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08352-3

Keywords: sarcopenia, frailty, exercise prescription, network meta-analysis, resistance training, elastic band training, machine-based training, multicomponent training, handgrip strength, functional mobility, older adults, dose-response