Cognitive Dysfunction and Other Long-Term Complications of Surgery and Anesthesia

Preexisting Cognitive Impairment

PreCI was used to refer to objectively assessed cognitive impairment that is observed in patients at baseline (compared to population norms). This is a preoperative assessment of impairment and should be considered in terms of cognitive impairment that might be coincidentally identified in the community and not just in terms of impending anesthesia and surgery. Hence, it is recommended that the term PreCI be replaced by mild NCD (MCI) or major NCD (dementia).

Delirium

POD should be recognized as a specific category consistent with DSM-5 terminology if the patient is in the immediate postoperative period and other specific causes have been excluded. The reported incidence of POD in the elderly is highly dependent on how it is diagnosed and screened. The most widely used and validated tool is the Confusion Assessment Method (CAM) ( Box 82.1 ) . The term postoperative refers to a specific and known temporal association with anesthesia and surgery, noting that surgical procedures occur annually in approximately 30% of individuals aged 65 years or more. POD is therefore defined as delirium which occurs in hospital up to 1 week postprocedure or until discharge, and which meets DSM-5 diagnostic criteria.

Postoperative Cognitive Dysfunction

POCD has been used in research studies to describe an objectively measurable decline in cognitive function at intervals from one day to 7.5 years after surgery. Significant heterogeneity around the definitions, time points for assessment, and criteria for POCD has resulted in widely varying results. As stated above, POCD does not require a subjective complaint or evidence of functional impairment, while the DSM-5 requires both. Thus, the major differences between POCD and NCD are the requirement for a cognitive concern, evidence of daily function, and the requirement for an objective decline in only one cognitive domain in the latter.

Cognitive Concern

The subjective cognitive complaint could be reported by the individual, family member, caregiver, or clinician. It is unlikely in the early postoperative period that a patient or an informant would be able to make an accurate assessment of subtle cognitive decline. Therefore, while assessment for NCD after discharge but prior to full recovery may be technically possible, the clinical relevance of its attribution would be unclear. Therefore, the term “delayed neurocognitive recovery” (dNCR) is recommended for this interval. This term should be used up to 30 days after the procedure, when recovery from most surgery and hospitalization should have occurred. The diagnostic criteria for NCD may be applied, but the outcome if impaired would be dNCR. It is possible that high functioning individuals may report cognitive concerns without objective evidence of decline. Because the subjective report from the participant, informant, or clinician is an essential element of diagnosing a PND, this may still be considered dNCR, and considered in the context of each individual case for clinical interpretation.

Activities of Daily Living Assessment

Assessment of daily function is an essential element of classifying mild and major NCD. This is achieved using an appropriate tool to measure ADL, which are everyday personal care activities that are fundamental to caring for oneself and maintaining independence. To detect more subtle declines in function, Instrumental Activities of Daily Living (IADLs) are used, which include activities like shopping, driving, and managing finances. For mild NCD (MCI), ADLs are maintained, while for major NCD (dementia), a decline is required.

Objective Testing

According to DSM-5 ⁵ and our recommended PND terms, mild NCD (postoperative) requires a 1 to <2 SD decline in cognitive testing compared to controls or norms, and major NCD (postoperative) requires a decline of 2 or more standard deviations on one or more cognitive domains (complex attention, executive function, learning and memory, language, perceptual-motor, or social cognition) using an appropriate neuropsychological assessment. Neither the DSM-5 nor the NIA-AA criteria for objective testing specify individual neuropsychological tests, nor the number of tests required in a battery. It is important to note that this refers to psychometric assessments that objectively assess specific cognitive domains, not to the use of screening tools such as the Mini-Mental State Examination (MMSE) or the Montreal Cognitive Assessment (MoCA). While useful for preoperative risk assessment, these brief screening tools do not have the sensitivity to assess change in specific cognitive domains.

Thirty days after surgery, the term mild NCD (postoperative) (postoperative MCI) or major NCD (postoperative) (postoperative dementia) should be used instead of dNCR. The “postoperative” modifier should be used as long as the criteria are met, and as long as first diagnosed prior to 12 months postoperatively. If first noticed or diagnosed 12 or more months following surgery, the postoperative modifier is not used.

Assumptions Underlying Assessment of Postoperative Cognitive Dysfunction/Perioperative Neurocognitive Disorder

Prospective studies investigating POCD usually assessed baseline cognition at a single time point only, often within days to weeks of the procedure. This makes three important and not necessarily valid assumptions. First, it assumes normal preoperative cognitive function, because screening was performed with a tool insensitive to cognitive impairment, such as the MMSE or MoCA. Second, a single time-point assessment assumes stable cognition. The third assumption is that results are reproducible, even without an intervention. Even if we assume that patients have stable cognition, it is unlikely that two consecutive assessments would yield the same results, due to a variety of intrinsic and extrinsic factors.

Cognitive Decline Outcome Criteria

There are several criteria that studies have used to define POCD in the past. These include the 1SD rule, the 20% rule, and the reliable change index (RCI). Each has advantages and limitations, but the latter has the large advantage of incorporating the changes observed in a control group over a similar time period, largely to account for practice and time effects. Because the RCI can be related to a control group or normative data, it satisfies the DSM-5 criteria for NCD (postoperative).

Sensitivity

The number of tests directly impacts the sensitivity of the measure; for example, if the definition of POCD is a decline of 2SD in 2 tests from a battery of 8 tests, the probability of identifying a decline of 2SD in 2 tests from 10 tests is doubled to 0.10. For the assessment of NCD, the DSM-5 only requires decline in one or more cognitive domains (complex attention, executive function, learning and memory, language, perceptual-motor, or social cognition), but does not recommend specific neuropsychological tests or a specific number of tests. Sensitivity can be improved by considering small declines across a number of tests, or a single very large decline in what is termed, “combined z-scores.”

While seemingly esoteric, the various methods of analysis can markedly impact the diagnosis of POCD/PND, highlighted by Keizer et al., who reported a POCD incidence of 10.5%, 31%, and 7.7% using the same test results with different methods of analysis (1SD definition, 20% definition and RCI, respectively). Clearly, standardization will be required to interpret results across many studies.

More recently, computerized cognitive assessment batteries have been developed that avoid practice effects, offer easier administration, are quicker and standardized, and have the potential to overcome cultural and language difficulties. To date, however, computerized test batteries have received limited attention in POCD/PND research mainly due to a lack of relevant validation studies.

group change versus individual change

Studies investigating POCD have variously used either individual or group change. Individual change refers to a dichotomous outcome of “decline” versus “no decline,” whereas group analysis considers differences between groups on a continuous scale. Clinical studies typically then use statistics to compare an a priori primary outcome between two groups to test whether any observed difference is due to chance. But this approach misses important individual information within each group. In particular, the individuals who decline the most may be a small fraction of the group, but arguably are the most critical to consider. Some individuals may even demonstrate cognitive improvement, due to correction of underlying pathology and improved function (e.g., reduction in pain, improved mobility, improved ADL). Statistics that simply average these groups together will miss these important individual changes ( Fig. 82.1 ), which might be due to identifiable and perhaps correctable factors.

Issues Associated With Repeated Testing

Serial neuropsychological testing is important to reliably assess the trend over time; however, repeated testing may introduce several sources of error, such as reliability, floor/ceiling effects, and practice effects. It is unclear if practice effects can be eliminated, as such effects can be observed as far out as 2.5 years. Individual (e.g., age, sex, culture, language, education, comorbidities, baseline cognitive function) as well as perioperative factors (anxiety, medications, pain, etc.) may also modulate the effect of repeated testing. Other strategies to improve reliability include parallel test versions and the use of control groups ( vide supra ). The major concern with control groups is selecting a well-matched one. It is generally desirable to match for comorbid conditions, even including the need for surgery, to evaluate the effect of the surgery itself.