Cognition After Postoperative Delirium—The Confusion Has Cleared but All Is Not Well

Postoperative delirium is common in older adults, but some of its long-term consequences remain debated. Its association with adverse outcomes is well established, including longer hospital stays, functional decline, institutionalization, and mortality. A more challenging question is whether delirium itself contributes to later cognitive decline or if it mainly identifies patients who already have frailty, have a health record that is medically complex, and are likely to become ill again. These competing explanations have distinct implications for care, including what we tell patients and families at discharge, how we plan follow-up, and how much emphasis we place on delirium prevention. In this issue of JAMA Internal Medicine, Hshieh et al¹ provide data that help advance this debate.

Their work used data from the SAGES study, which included 560 older adults without clinical dementia at baseline who underwent elective surgery and were followed up with repeated neuropsychological testing over 5 years.¹ That extended follow-up, uncommon in delirium studies, allowed researchers to examine cognitive trajectories and subsequent health events beyond the immediate postoperative period. Patients who experience delirium are typically older and have more severe illness than those who do not. They are also more likely to be rehospitalized, need intensive care, or require postacute care after discharge. Any of these events could worsen cognition. The key issue, then, is whether postoperative delirium is a marker of illness and frailty in patients who are prone to repeated hospitalizations and cognitive decline or if there is something about the delirium itself that changes the long-term course of cognitive function.

The findings did not support the first explanation. Participants who developed postoperative delirium were rehospitalized more often, and rehospitalization was itself associated with worse short-term cognitive outcomes. However, when the authors¹ modeled rehospitalization over time as a possible mediator (ie, as something that might explain how delirium actually connects to later decline), the association between delirium and long-term cognitive decline remained largely unchanged. Similar results were observed when examining rehospitalizations involving intensive care unit stays or postacute care separately. Overall, the findings indicated that repeated hospitalizations contributed to cognitive decline but did not account for the worse long-term cognitive trajectories associated with delirium.

A helpful way to understand the results is to distinguish between temporary cognitive setbacks and a long-term change in cognition. Rehospitalizations were associated with cognitive losses, which could build up with each event. However, these losses were relatively short-lived and occurred within the context of the patient’s broader decline over time. On the other hand, delirium showed a different pattern. It was associated with persistent deterioration. This distinction is important because it suggests that, even though delirium may seem temporary during a hospital stay, it is associated with a lasting change in the patient’s cognition.

Admittedly, these findings need to be understood within the context of the study’s limitations. Hospitalization is an imperfect measure of illness and frailty that does not capture a wider range of factors, such as outpatient exacerbations, pain, medication effects, nutritional problems, sensory deprivation, immobility, sleep disruption, depression, or social stressors, all of which can also affect recovery and cognition. These factors can plausibly affect cognition without resulting in readmission. Therefore, the absence of mediation through rehospitalization does not rule out an important role for underlying illness and frailty; it may simply mean that rehospitalization was too narrow a measure of these domains. The analysis also could not account for recurrent delirium after discharge, which may matter if long-term decline stems from repeated insults to a vulnerable brain. Another consideration is generalizability, since the cohort was predominantly White, highly educated, and from 2 tertiary care centers.

Even so, the study had notable strengths, building on earlier work from the same cohort. These included prospective identification of incident delirium, careful triangulation of rehospitalizations, repeated cognitive testing, and long-term follow-up with attention to attrition. In 2023, Kunicki et al² reported that postoperative delirium was associated with a 40% faster rate of long-term cognitive decline over 6 years after elective surgery. The new analyses went further, indicating that delirium might be more than just a cognitive marker of illness and frailty. Moreover, the results aligned with a growing body of evidence. A meta-analysis found a consistent association between delirium and later cognitive decline across surgical and nonsurgical settings, and neuropathologic studies suggest that delirium may interact with, rather than merely reveal, underlying dementia pathology.^3,4 In this context, the authors help move the field forward by showing that hospitalization-related events contribute to cognitive decline but account for little of the persistent association between delirium and worse long-term cognition.

The findings are especially relevant in the perioperative setting, in which there is a clearer opportunity to intervene. Much of the burden of delirium occurs in medical patients, many of whom already have delirium when admitted or develop it early during an acute illness.^5,6 In those situations, prevention may not be possible, and the causal pathway may be more difficult to clarify. Postoperative delirium is different. The risk window is expected, preoperative risk factors can be identified, precipitating factors can be modified, and preventive strategies can be implemented. Multicomponent nonpharmacologic interventions have been proven to reduce the incidence of delirium and remain among the most practical and evidence-based methods for protecting brain health in hospitalized older adults.⁷ If delirium can directly lead to cognitive decline, these findings further strengthen the case for using those strategies consistently.

There are additional clinical implications. Rehospitalization is not harmless and was clearly associated with short-term cognitive decline, which reinforces the importance of reducing readmissions, improving care transitions, promoting early mobilization, simplifying medication regimens, and paying close attention to postacute care. Postoperative delirium should not be treated as a problem that ends once the confusion improves. It should prompt clinicians to prepare families for a recovery that may be prolonged and uneven and consider follow-up plans that include monitoring cognition and function after discharge. The traditional distinction between reversible delirium and progressive dementia may still be useful at the bedside, but it does not fully capture the long-term cognitive consequences associated with delirium.

For researchers and health systems, the main question should be less about whether delirium deserves long-term attention and more about how to respond effectively once it occurs. Prevention remains vital, but it may not be sufficient. If delirium marks a turning point in brain health, future studies should explore whether the care delivered during the episode, as well as in the following weeks and months, can change later cognitive outcomes. Pragmatic trials embedded in routine perioperative and hospital care may be especially valuable because the same features that make delirium clinically significant (eg, multifactorial causes, association with function and caregiving) also make it difficult to study in traditional clinical trials.

Older adults who develop delirium are undoubtedly vulnerable, and efforts to prevent avoidable rehospitalizations remain crucial. Nonetheless, the findings of Hshieh et al¹ indicate that the delirium episode itself has prognostic significance that is only minimally explained by underlying illness and frailty. Therefore, postoperative delirium should be regarded as a complication with the potential for long-term consequences. An episode that seems to resolve during hospitalization may still matter years later. This perspective should influence not only how clinicians advise patients and care partners but also how health systems prioritize delirium prevention and posthospital recovery.