Postoperative Cognitive Complications

Abstract

Postoperative cognitive complications, a form of cognitive dysfunction, are declines in memory, attention, and executive function that follow surgery and anaesthesia, most often in older patients. The field once gathered every such change under the single name postoperative cognitive dysfunction, but a 2018 consensus divided it by timing into postoperative delirium, delayed neurocognitive recovery, and postoperative neurocognitive disorder. The deficits are usually mild and, for most patients, ease over the months after surgery, though a minority are left with a lasting impairment. Because the diagnosis rests on a measured change from each patient's own pre-surgical baseline, judged against how much healthy controls change on retesting, the condition is as much a problem of measurement as of medicine. This article sets out the entities, their mechanisms and risk factors, their course, and the psychometrics of detecting decline, with three interactive demonstrations.

Keywords: postoperative cognitive complications, postoperative cognitive dysfunction, perioperative neurocognitive disorders

Postoperative cognitive complications are cognitive declines associated with surgery and the anaesthesia that accompanies it, ranging from an acute confusional state in the first days to a measurable shortfall in memory and executive control that can persist for months (Evered et al., 2018). In the Medical Subject Headings vocabulary the descriptor is Postoperative Cognitive Complications, filed at once among the cognition disorders and among the postoperative complications, so the entity is itself catalogued as a disorder rather than a passing symptom. Its principal entry term is the label under which most of the research literature was written, postoperative cognitive dysfunction, abbreviated POCD. No dedicated ICD-10 or ICD-11 code exists for the condition; in clinical records it is captured under broader headings for cognitive or postoperative disturbance, and the precise vocabulary that names it is the MeSH descriptor together with the 2018 consensus nomenclature. What makes the condition instructive for cognitive psychology is that it forces the discipline's measurement questions into the open: what counts as a decline when everyone's scores drift on retesting, how the timing of a deficit changes what it is called, and why age matters so much more than the choice of anaesthetic.

What Postoperative Cognitive Complications Are

Patients and clinicians have long noticed that some people are not quite themselves after an operation, and that the elderly are especially vulnerable to a period of confusion, forgetfulness, and mental slowing that outlasts the anaesthetic (Moller et al., 1998). Postoperative cognitive complications name the measurable version of that observation: changes in cognition, detected on testing against a pre-surgical baseline, that follow surgery and anaesthesia. The changes are usually modest. On formal testing the deficits are small, a fraction of a standard deviation, and they concentrate on a handful of domains, chiefly memory, attention, executive function, and processing speed, rather than dulling cognition as a whole. This is the first thing to understand, and the thing most easily lost when a frightening operation and a frightening word like dysfunction meet: for the great majority of patients the change is mild and, in time, largely reversible. What complicates the picture is timing. A confusional state on the first postoperative night and a subtle memory deficit six months later look nothing alike, arise through partly different mechanisms, and carry different implications, yet both were once filed under the same heading. Figure 1 places the complications on a timeline, from the surgical insult through the acute phase to the divergent trajectories of the following year.

The Perioperative Neurocognitive Disorders

For most of the field's history a single phrase, postoperative cognitive dysfunction, did all the work, and it did it badly, because it forced an acute delirium and a chronic memory deficit into one box (Evered et al., 2018). The 2018 recommendations for the nomenclature of cognitive change associated with anaesthesia and surgery repaired this by aligning the perioperative terms with the wider psychiatric classification of neurocognitive disorders and, crucially, by sorting them along the axis that matters most, the time since surgery. Three entities result. Postoperative delirium is the acute, fluctuating disturbance of attention and awareness that appears in the first hours to days, the most visible complication and the one most clearly linked to worse outcomes (Inouye et al., 2014). Delayed neurocognitive recovery is a decline from baseline detected on testing within the first 30 days, once any delirium has cleared, and it signals incomplete recovery rather than a settled disorder. Postoperative neurocognitive disorder is decline that persists between 30 days and 12 months, graded mild or major by whether it compromises independence; this is the entity that the older literature studied as POCD. Beyond 12 months, by convention, decline is no longer attributed specifically to the operation. The scheme is more than tidy bookkeeping: because the three entities differ in mechanism, risk, and prognosis, naming them apart is the precondition for studying them properly (Berger et al., 2015). Table 1 lays out the entities and their time windows, and the demonstration below lets the timing be set and names the entity that applies.

Table 1

The Time-Defined Perioperative Neurocognitive Entities of the 2018 Nomenclature

Note. The entities are distinguished chiefly by how long after surgery the cognitive change is present, aligning perioperative terminology with the psychiatric classification of neurocognitive disorders (Evered et al., 2018).

Mechanisms and Risk Factors

Two questions run through the science of postoperative cognitive complications, and they have different answers: what damages cognition, and who is vulnerable to it (Berger et al., 2015). On mechanism, the most productive shift has been to move the suspicion away from the anaesthetic drugs and toward the body's response to the surgery itself. A landmark line of work showed that peripheral surgical trauma triggers an inflammatory cascade, with cytokines such as tumour necrosis factor driving neuroinflammation that disrupts hippocampal function and memory (Terrando et al., 2010). This reframes the problem: the brain is affected less by the anaesthetic gas than by the systemic inflammatory consequence of being cut, which is why the nature and magnitude of the surgery matter and why the choice of anaesthetic matters less than intuition suggests. That counterintuitive fact is one of the field's firmest findings. A randomised comparison of regional against general anaesthesia in elderly patients found no meaningful difference in lasting cognitive outcome, undercutting the natural assumption that it is the general anaesthetic that clouds the mind (Rasmussen et al., 2003). On vulnerability, by contrast, one factor dominates all others: age. The large prospective cohorts converge on advancing age as the most consistent independent predictor of persistent postoperative decline, with additional contributions from lower education, prior cognitive impairment, and the burden of the surgery and its complications (Monk et al., 2008; Moller et al., 1998). Age and education together behave much like a cognitive reserve, the same buffering capacity invoked across the study of cognitive aging: a comparable insult surfaces as clinical impairment in the patient with less reserve and is absorbed by the patient with more. The demonstration below renders the age gradient, showing how the modelled risk of lasting decline climbs with the patient's age.

Course, Recovery, and Persistence

The single most reassuring fact about postoperative cognitive complications, and the one the numbers make plain, is that for most patients they resolve (Moller et al., 1998). The first International Study of Postoperative Cognitive Dysfunction, testing elderly patients before major non-cardiac surgery and again afterwards, found cognitive dysfunction in about a quarter of them at one week but in under a tenth at three months, the incidence falling steeply as the weeks passed. The trajectory is the familiar one: a deficit largest around the time of surgery that recovers substantially over the following months. But recovery is often partial, and a meaningful minority are left with a deficit that has not resolved, a persistence documented most strikingly in cardiac surgery, where a longitudinal study of coronary-artery bypass patients found cognitive decline that, having improved by discharge, re-emerged and was detectable five years later (Newman et al., 2001). The persistence is not merely a matter of test scores. Postoperative decline, and postoperative delirium in particular, is associated with worse long-term outcomes, including greater mortality and a raised risk of later dementia, so the complication is a marker of vulnerability as well as a burden in its own right (Steinmetz et al., 2009; Needham et al., 2017). Whether surgery and anaesthesia can actually cause or accelerate a dementing process, or whether they merely unmask a decline already under way in a vulnerable brain, remains among the field's open and consequential questions (Needham et al., 2017). The picture that emerges is neither the harmless transient dismissed by some nor the inevitable ruin feared by patients, but a genuine, mostly recoverable complication with a persistent and prognostically important tail.

Defining and Measuring: The Reliable Change Index

For a condition whose effects are small and whose very existence was once doubted, measurement is not a technical footnote but the substance of the science (Moller et al., 1998). The central difficulty is that a single postoperative test score means little, because cognition is not measured against a fixed standard but against the patient's own former self, and because scores move on their own when a test is taken twice. Healthy people who take a cognitive test and then retake it usually do a little better the second time, a practice effect, so a patient whose score merely stays flat has arguably declined relative to where practice should have carried them. The reliable change index was adopted to handle exactly this (Rasmussen et al., 2003). It expresses each patient's change from baseline as a standardised score within the distribution of change observed in a matched control group, and it counts the change as a genuine decline only when it passes a criterion, conventionally about 1.96 standard deviations in the impaired direction. Because the control group's mean change is typically a small improvement, the threshold a patient must cross to be called impaired is not zero but a decline large enough to overcome the expected practice gain. Applied across a battery of tests, the individual reliable-change scores are often combined into a single index, so that a patient is classified as impaired on the pattern across tests rather than on any one (Evered & Silbert, 2018). None of this is peculiar to surgery; it is the ordinary and demanding psychometrics of detecting change in an individual, which postoperative cognitive complications happen to lay bare. The demonstration below makes the logic visible, placing a patient's change against the control distribution and marking the line beyond which the change counts as reliable.

Worked Example

The reliable change index turns on a comparison that is worth making concrete, because it shows why a raw drop in score is not the same as a decline and how a battery of tests is combined into a single verdict. Suppose a control group of patients who did not have the cognitive-sensitive surgery, retested over the same interval, improved on a memory test by a mean of 2 points, with a standard deviation of change of 5 points; the small improvement is the practice effect. A surgical patient's score on that test falls by 10 points, a change of minus 10. The reliable-change z-score is the patient's change minus the control mean change, divided by the control standard deviation: minus 10 minus 2, all over 5, which is minus 12 over 5, that is minus 2.4. Because minus 2.4 is beyond the criterion of minus 1.96, the change counts as a reliable decline on this test. Notice what the practice correction does: the point at which a change becomes reliable is not a fall of zero but a fall past 2 minus 1.96 times 5, that is minus 7.8 points, so a patient who dropped only 6 points would not qualify despite a real-looking decline, because healthy controls of the same interval improved. Now combine tests. Suppose the same patient also declined on an executive-function test by 4 points where controls had improved 1 point with a change standard deviation of 4; that z-score is minus 4 minus 1, over 4, which is minus 1.25, short of the single-test criterion. The International Study convention combines the per-test z-scores into one by summing them and dividing by the square root of the number of tests: minus 2.4 plus minus 1.25 is minus 3.65, divided by the square root of 2, about 1.414, giving a combined z of about minus 2.58. That combined score is beyond minus 1.96, so the patient is classified with postoperative cognitive decline on the pattern across the battery, even though only one of the two tests reached the threshold alone. The arithmetic makes two things plain. First, decline is defined relative to expected change, not to zero, so the practice effect raises the bar. Second, aggregating across tests both stabilises the judgment and can detect a coherent decline that no single test would confirm, which is exactly why the field measures a battery rather than a single score.

Discussion

Postoperative cognitive complications earn their place in cognitive psychology less as a surgical side effect than as a natural experiment in the measurement of change within the individual. To study them at all is to confront the hardest problem in applied psychometrics: deciding whether one person's score has really moved, when the only yardstick is that person's own baseline and when scores drift on retesting from practice alone (Moller et al., 1998; Rasmussen et al., 2003). The reliable change index and its combination across a battery are the discipline's principled answer, and the condition is a clean setting in which to see them work (Evered & Silbert, 2018). The mechanism story is equally instructive, because it overturned an intuition: the damage tracks the inflammatory response to surgery more than the anaesthetic itself, which is why the choice of anaesthetic barely moves the outcome while the magnitude of the surgery and the vulnerability of the patient do (Terrando et al., 2010; Rasmussen et al., 2003). The dominance of age as a risk factor connects the condition to the wider theory of cognitive reserve, framing the wide spread of outcomes as different starting points meeting a common insult (Monk et al., 2008). And the reorganisation of the terminology into time-defined entities is a small case study in how a field sharpens its object of study by naming its parts, separating an acute delirium from a chronic disorder that had been blurred together for decades (Evered et al., 2018). The honest summary for the patient and the accurate one for the scientist coincide: a real but usually mild complication, most often recoverable, more a marker of a vulnerable brain than a poisoning of a healthy one, and best understood through the careful measurement of change (Needham et al., 2017).

Common Misconceptions

The general anaesthetic is what damages the brain.

A randomised comparison of regional against general anaesthesia in elderly patients found no meaningful difference in lasting cognitive outcome. The evidence points to the inflammatory response to the surgery itself, not the anaesthetic gas, as the main driver of persistent decline (Rasmussen et al., 2003; Terrando et al., 2010).

Any drop in test score after surgery is postoperative cognitive decline.

Decline is a reliable change from the patient's baseline, judged against how much a control group's scores move on retesting. Because controls usually improve a little from practice, a small fall does not qualify; the change must be large enough to clear the practice-corrected criterion (Rasmussen et al., 2003; Evered & Silbert, 2018).

It is always permanent, or always temporary.

Neither. For most patients the deficit is largest around surgery and recovers over the following months, yet a minority are left with a persistent impairment, and postoperative decline is associated with worse long-term outcomes. The course is recoverable for the majority but has a real and prognostically important tail (Moller et al., 1998; Steinmetz et al., 2009).

Glossary

Anaesthesia.

The controlled, reversible loss of sensation or consciousness for surgery; long suspected as a cause of postoperative cognitive complications, though the evidence now implicates the surgery more than the anaesthetic.

Attention.

The selection and sustaining of focus on relevant information; its acute disturbance is the core feature of postoperative delirium.

Cognitive reserve.

The brain's capacity to sustain function despite insult, built from education and prior ability; higher reserve helps explain why the same surgery spares one patient and impairs another.

Cytokine.

A signalling protein of the immune system; the surgical wound releases cytokines that carry an inflammatory signal to the brain, the first step in the neuroinflammation account of postoperative cognitive decline.

Delayed neurocognitive recovery.

Cognitive decline from baseline detected on testing within the first 30 days after surgery, once any delirium has resolved; the term for incomplete early recovery.

Executive function.

The control processes of planning, switching, and inhibition; among the domains that decline in postoperative neurocognitive disorder.

Neuroinflammation.

An inflammatory response within the nervous system; triggered by surgical trauma through a cytokine cascade, it is a leading candidate mechanism for postoperative cognitive decline.

Perioperative neurocognitive disorders.

The umbrella term of the 2018 nomenclature covering the time-defined cognitive entities associated with surgery and anaesthesia.

Postoperative delirium.

An acute, fluctuating disturbance of attention and awareness in the first hours to days after surgery; the earliest perioperative neurocognitive event.

Postoperative neurocognitive disorder.

Cognitive decline from baseline persisting between 30 days and 12 months after surgery, graded mild or major; the entity historically studied as POCD.

Practice effect.

The tendency to score better on a second exposure to a cognitive test; the reason a patient whose score merely holds steady may have declined relative to expectation.

Processing speed.

The rate at which mental operations are carried out; among the domains most sensitive to postoperative cognitive change.

Prospective study.

A design that tests patients from a pre-surgical baseline forward, allowing the effect of surgery to be separated from pre-existing differences.

Reliable change index.

A statistic expressing an individual's change from baseline as a standardised score in a control group's change distribution, used to decide whether a change is a genuine decline rather than retest noise.

Key Researchers

Roderic G. Eckenhoff. Anaesthesiologist and Austin Lamont Professor at the University of Pennsylvania; his laboratory studies the molecular actions of anaesthetics and their relationship to neurodegeneration, informing the debate over whether surgery and anaesthesia can accelerate cognitive decline. Faculty Page - ORCID

Lisbeth A. Evered. Perioperative-neuroscience researcher at the University of Melbourne and Weill Cornell Medicine; she led the 2018 recommendations that recast the terminology of cognitive change after anaesthesia and surgery into time-defined entities. Faculty Page - ORCID - Google Scholar

Sharon K. Inouye. Geriatrician at Harvard Medical School and the Marcus Institute for Aging Research; she developed the Confusion Assessment Method and is a leading authority on delirium in older adults, the acute end of the perioperative neurocognitive spectrum. Faculty Page - Google Scholar - Wikipedia

Lars S. Rasmussen. Anaesthesiologist and professor at the University of Copenhagen and Rigshospitalet; he led the International Study of Postoperative Cognitive Dysfunction, whose prospective cohorts established the incidence, risk factors, and measurement of postoperative cognitive decline. Faculty Page - ORCID

Brendan S. Silbert. Anaesthetist at St Vincent's Hospital Melbourne and the University of Melbourne; he co-authored the 2018 nomenclature and reviews of postoperative cognitive dysfunction in noncardiac surgery, and studies its detection and course. Faculty Page - ORCID - Google Scholar

Frequently Asked Questions

What are postoperative cognitive complications?
They are cognitive declines associated with surgery and anaesthesia, affecting memory, attention, executive function, and processing speed. In the 2018 nomenclature they range from acute postoperative delirium to a postoperative neurocognitive disorder lasting up to a year (Evered et al., 2018).

Is it caused by the anaesthetic?
Largely not. A randomised comparison of regional against general anaesthesia found no meaningful difference in lasting cognitive outcome, and the leading mechanism is the inflammatory response to the surgery itself rather than the anaesthetic drug (Rasmussen et al., 2003; Terrando et al., 2010).

Who is most at risk?
Older patients, above all. Advancing age is the most consistent independent predictor of persistent decline across the large cohort studies, with lower education, prior cognitive impairment, and the burden of the surgery adding to the risk (Monk et al., 2008; Moller et al., 1998).

How is a decline actually diagnosed?
By a reliable change from the patient's own pre-surgical baseline, not a single low score. The change is compared against how much a control group's scores move on retesting, and counts as a decline only when it passes a practice-corrected criterion (Rasmussen et al., 2003; Evered & Silbert, 2018).

What is the difference between delirium and postoperative cognitive dysfunction?
Postoperative delirium is acute and fluctuating, appearing in the first days, whereas the disorder once called POCD is a slower, more stable decline detected on testing weeks to months later. The 2018 nomenclature separates them precisely because they differ in timing, mechanism, and prognosis (Evered et al., 2018; Inouye et al., 2014).

Does it get better?
For most patients, yes. Longitudinal studies find the incidence falling steeply over the first months, from about a quarter of elderly patients at one week to under a tenth by three months, though a minority retain a persistent deficit (Moller et al., 1998).

Is it linked to dementia?
Postoperative decline, and delirium in particular, is associated with worse long-term outcomes, including a raised risk of later dementia. Whether surgery accelerates a dementing process or merely unmasks one already under way is an open question (Steinmetz et al., 2009; Needham et al., 2017).

Is there an ICD code for it?
There is no dedicated ICD-10 or ICD-11 code; in records the condition is captured under broader headings for cognitive or postoperative disturbance. The MeSH descriptor Postoperative Cognitive Complications and the 2018 consensus nomenclature are the vocabularies that name it most precisely (Evered et al., 2018).

Support Organizations

Organizations that provide information and guidance on surgery, anaesthesia, and cognitive health in older patients.

References