Assessing patterns of delirium symptoms reveals a novel subtype among elective surgical patients with postoperative delirium

Abstract

Objectives: Prior studies reported incidence of hypoactive and hyperactive subtypes of postoperative delirium, but did not consider cognitive symptoms of delirium which are highlighted in the DSM-5 criteria for delirium. This study aims to address this gap in the literature by classifying cases of delirium according to their constellation of cognitive and motoric symptoms of delirium using a statistical technique called Latent Class Analysis (LCA).

Methods: Data were from five independent study cohorts (N = 1968) of patients who underwent elective spine, knee/hip, or elective gastrointestinal and thoracic procedures, between 2001 and 2017. Assessments of delirium symptoms were conducted using the long form of the Confusion Assessment Method (CAM) pre- and post-surgery. Latent class analyses of CAM data from the first 2 days after surgery were conducted to determine subtypes of delirium based on patterns of cognitive and motoric symptoms of delirium. We also determined perioperative patient characteristics associated with each latent class of delirium and assessed whether the length of delirium for each of the patterns of delirium symptoms identified by the latent class analysis.

Results: The latent class model from postoperative day 1 revealed three distinct patterns of delirium symptoms. One pattern of symptoms, denoted as the Hyperalert class, included patients whose predominant symptoms were being hyperalert or overly sensitive to environmental stimuli and having a low level of motor activity. Another pattern of symptoms, denoted as the Hypoalert class, included patients whose predominant symptom was being hypoalert (lethargic or drowsy). A third pattern of symptoms, denoted as the Cognitive Changes class, included patients who experienced new onset of disorganized thinking, memory impairment, and disorientation. Among 352 patients who met CAM criteria for delirium on postoperative day 1, 34% had symptoms that fit within the Hyperalert latent class, 39% had symptoms that fit within the Hypoalert latent class, and 27% had symptoms that fit within the Cognitive Changes latent class. Similar findings were found when latent class analysis was applied to those who met CAM criteria for delirium on postoperative day 2. Multinomial regression analyses revealed that ASA class, surgery type, and preoperative cognitive status as measured by the Telephone Interview for Cognitive Status (TICS) scores were associated with class membership. Length of delirium differed between the latent classes with the Cognitive Changes latent class having a longer duration compared to the other two classes.

Conclusions: Older elective surgery patients who did not have acute events or illnesses or a diagnosis of dementia prior to surgery displayed varying symptoms of delirium after surgery. Compared to prior studies that described hypoactive and hyperactive subtypes of delirium, we identified a novel subtype of delirium that reflects cognitive symptoms of delirium. The three subtypes of delirium reveal distinct patterns of delirium symptoms which provide insight into varying risks and care needs of patients with delirium, indicating the necessity of future research on reducing risk for cognitive symptoms of delirium.

Keywords: aging; delirium subtypes; elective surgery; postoperative delirium.