Hawaii Department of Health data consistently identify postoperative delirium as one of the more consequential yet underreported complications affecting older surgical patients across the state’s hospital system. Against that backdrop, a systematic review and network meta-analysis published in The BMJ on February 12, 2026, offers a structured examination of pharmacological interventions aimed at reducing delirium incidence in this population. The study, authored by Matthew Luney and colleagues (BMJ 2026;392:e085539, doi:10.1136/bmj-2025-085539), synthesizes randomized controlled trial (RCT) data to evaluate the comparative effectiveness of drug-based prevention strategies for postoperative delirium (POD) in older adults.

Postoperative Delirium: A Population-Level Concern

Postoperative delirium represents a clinically significant complication in geriatric surgical populations, characterized by acute disturbances in attention, awareness, and cognition that develop over a short period and fluctuate throughout the day. Prevalence estimates vary by surgical context, but published surveillance data indicate POD occurs in approximately 15 to 25 percent of older adults following major noncardiac surgery, with rates exceeding 50 percent in high-risk subgroups such as those undergoing cardiac or orthopedic procedures.

The public health burden is considerable. POD is associated with prolonged hospital stays, accelerated cognitive decline, increased rates of institutionalization, and elevated short-term mortality. For Hawaii’s rapidly aging population, which the State Department of Health projects will see the proportion of residents aged 65 and older rise to nearly 22 percent by 2030, the prevalence of POD carries direct implications for hospital capacity, post-acute care utilization, and long-term care infrastructure.

Despite these documented burdens, pharmacological prevention strategies for POD have not achieved consensus in clinical practice. Multiple drug classes have been evaluated across individual RCTs, including antipsychotics, melatonin receptor agonists, alpha-2 adrenergic agonists, ketamine, and cholinesterase inhibitors, but heterogeneity in trial design, patient populations, and outcome definitions has historically complicated direct comparisons.

Study Design and Methodological Approach

The Luney et al. network meta-analysis (NMA) addresses this comparative gap by applying an NMA framework to data drawn from eligible RCTs. Network meta-analysis allows for indirect comparisons between interventions that have not been tested head-to-head within a single trial, synthesizing evidence across a connected web of studies. This methodology is particularly valuable in the POD prevention literature, where most trials compare a single active agent against placebo rather than against one another directly.

The review adhered to established systematic review methodology. Eligible studies enrolled adult participants, the majority of whom were older adults, who were randomized to receive a pharmacological intervention or control condition in the perioperative period. The primary outcome of interest was the incidence of POD, typically defined according to validated diagnostic criteria such as the Confusion Assessment Method (CAM) or equivalent instruments. Secondary outcomes likely addressed delirium severity, duration, and associated complications, consistent with standard NMA reporting in this domain.

The authors conducted a correction to the published article, noting that co-author Arwa Hagana’s name had been misspelled in the original February 12, 2026 publication. The correction was issued to ensure accurate attribution across indexing databases and citation records, a procedurally standard but consequential step in maintaining the integrity of the scientific record.

Drug Classes Under Evaluation

The breadth of pharmacological agents examined in this NMA reflects the heterogeneous mechanistic hypotheses underlying POD pathophysiology. Current evidence suggests POD arises through multiple converging pathways, including neuroinflammation, cholinergic deficiency, dopaminergic excess, and disruption of melatonergic circadian signaling. Different drug classes target distinct aspects of this pathophysiology.

Antipsychotics, including haloperidol and atypical agents such as olanzapine and quetiapine, have historically received the most clinical attention given their dopamine antagonist properties. Melatonin and ramelteon have been evaluated for their capacity to stabilize sleep-wake cycles, disruption of which is both a precipitant and a manifestation of delirium. Dexmedetomidine, an alpha-2 adrenergic agonist with sedative and analgesic properties, has demonstrated notable results in intensive care unit populations and has been the subject of increasing investigation in the perioperative context. Ketamine, administered at subanesthetic doses, has been examined for its N-methyl-D-aspartate (NMDA) receptor antagonist effects on neuroinflammatory cascades. Cholinesterase inhibitors, such as rivastigmine, address the cholinergic hypothesis but have yielded inconsistent results in prior trials.

The NMA framework permits ranking of these agents according to their relative efficacy in reducing POD incidence, a functionality that individual RCTs cannot provide. Such rankings, while requiring cautious interpretation given the assumptions embedded in NMA methodology, offer clinicians a structured basis for comparative decision-making.

Relevance to Geriatric Surgical Practice in Hawaii

Hawaii’s hospital system serves a demographically distinctive surgical population. The state’s older adults include substantial proportions of Native Hawaiian, Japanese American, Filipino American, and other Asian and Pacific Islander individuals, population groups for whom pharmacogenomic variation may influence both drug metabolism and adverse effect profiles. Polymorphisms in cytochrome P450 enzymes, for instance, are known to affect the metabolism of haloperidol and certain other psychoactive agents at frequencies that differ across ancestral populations.

This consideration does not diminish the relevance of findings from the Luney et al. review but underscores the importance of contextualizing NMA results within local patient populations. Pharmacological interventions demonstrating efficacy at the population level in predominantly European or East Asian trial cohorts may require dosing adjustment or closer monitoring in Hawaii’s multiethnic surgical patient population.

Additionally, Hawaii’s geographic context introduces logistical considerations in perioperative care. Neighbor island hospitals, which serve patients on Maui, Hawaii Island, Kauai, and Molokai, operate with more constrained specialist resources than Oahu’s tertiary centers. Implementation of pharmacological POD prevention protocols at these facilities requires agents that can be managed safely without continuous neurological monitoring infrastructure that may not be uniformly available. Findings that identify effective and operationally feasible interventions are therefore of particular practical interest to hospital pharmacists, anesthesiologists, and geriatric medicine consultants working across the state system.

Interpreting Network Meta-Analysis Findings

Clinicians and public health practitioners interpreting NMA findings should attend to several methodological considerations that govern the strength of conclusions drawn from this evidence synthesis.

First, the validity of indirect comparisons in an NMA depends on the transitivity assumption, specifically, that patient populations and clinical contexts are sufficiently similar across the network of trials to permit meaningful comparison. Variability in surgical type, patient age distribution, baseline cognitive status, and delirium assessment methodology across included RCTs may introduce inconsistency that limits the precision of estimated treatment rankings.

Second, sample sizes within individual trial arms vary considerably across the POD pharmacological prevention literature. Smaller trials contribute estimates with wider confidence intervals, and the pooled NMA estimates should be evaluated with attention to heterogeneity statistics and credible intervals rather than point estimates alone.

Third, the outcome of POD incidence, while clinically meaningful, represents only one dimension of a complex syndrome. A drug may reduce the rate of delirium onset while exerting adverse effects on sedation depth, respiratory function, or hemodynamic stability that are not fully captured in delirium-focused outcome reporting. Evaluation of the benefit-risk profile of any pharmacological prevention strategy must incorporate safety data alongside efficacy estimates.

Implications for Clinical Protocol Development

For hospital systems considering the integration of pharmacological POD prevention into perioperative care pathways, the Luney et al. NMA provides a structured evidentiary foundation. The Queen’s Medical Center, Straub Medical Center, and Hawaii Pacific Health’s network, along with the Hawaii Health Systems Corporation’s neighbor island facilities, each maintain anesthesia and surgical quality improvement programs within which evidence-based delirium prevention protocols could be embedded.

Multidisciplinary perioperative teams, incorporating input from anesthesiology, geriatric medicine, pharmacy, and nursing, are best positioned to translate NMA findings into locally adapted protocols. The American Geriatrics Society’s Hospital Elder Life Program (HELP) model provides a nonpharmacological framework against which pharmacological strategies can be layered for patients at elevated delirium risk as assessed by validated preoperative screening instruments.

The Centers for Disease Control and Prevention (CDC) and the Agency for Healthcare Research and Quality (AHRQ) have each identified hospital-acquired delirium as a priority target for patient safety improvement. Systematic review evidence of the caliber produced by Luney and colleagues contributes directly to the evidentiary base supporting AHRQ’s safety program recommendations and may inform