A narrative review of preventive measures for postoperative delirium in older adults

Postoperative delirium (POD) is an acute neurological condition that develops as an adverse complication following surgery or use of anaesthesia. It usually lasts for some hours postoperatively but, if left untreated, it can manifest for several days (Bettelli and Neuner, 2017). Delirium can be characterised as hyperactive, hypoactive and mixed. Patients with hyperactive delirium will have symptoms of agitation, acute disorientation or restlessness. Hypoactive delirium commonly goes unrecognised because patients usually present as lethargic and quiet, but still disoriented (Flynn Makic, 2013).

Delirium has been well documented as one of the leading complications of major surgery, affecting 15–53% of patients; however, the incidence of POD varies and depends on age, risk factors and type of surgery (Gherghina et al, 2011; Inouye et al, 2014; Wang et al, 2018). It is a serious complication for older adults because an episode of delirium can start a cascade of adverse events, for example a prolonged length of stay in hospital but also increase the risk of mortality due to complications such as hospital-acquired infection, pressure ulcers, incontinence and falls (Wass et al, 2008; O'Neal and Shaw, 2016).

Despite the significance and severity of POD, it is still an underestimated phenomenon, and its early and regular detection and prevention is limited due to the low-quality evidence that contributes to current perioperative guidelines (Glynn and Corry, 2015; Vlisides and Avidan, 2019). Delirium is often misdiagnosed as depression or dementia, or considered normal behaviour in older patients due to lack of knowledge and awareness (Weng et al, 2019). A greater importance should be placed on proactive rather than reactive approaches in health care to minimise the chances of POD. Prevention may be the most effective strategy for reducing incidence and adverse outcomes; however, recent trials investigating the effect of drugs, such as ketamine and dexmedetomidine, have reported inconsistent findings and have suggested that non-pharmacological multicomponent initiatives may be more effective (Inouye et al, 2014; Vlisides and Avidan, 2019). Therefore, the purpose of this literature review is to explore prevention strategies for POD in older adults.

Search strategy

An initial search of two databases, CINAHL and PubMed, sourced 309 articles. Keywords included ‘delirium’, ‘acute confusion’, ‘intervention’, ‘nurs*, ‘patient’’, ‘preoperative’, ‘intraoperative’ ‘postoperative’.

Inclusion criteria

All studies that had recruited participants over the age of 65 in acute care settings post-surgery were included. The inclusion criteria were:

• Research studies that used quantitative or qualitative methodologies
• Systematic reviews
• Literature reviews
• Dissertations.

Studies were also included if they discussed preventive strategies, screening methods or interventions for reducing the incidence of POD during perioperative stages. Exclusion criteria were:

• Studies that discussed the treatment or management of POD without any clear mention of prevention strategies
• Electronically inaccessible
• Exceeded past 30 years
• If a study only examined non-surgical delirium, which was irrelevant to the focus of the review.

Limiters included articles from peer review journals and were written in English.

Following a screening of titles and abstracts, 25 studies met these inclusion criteria (Table 1). The generic framework for appraisal of research studies was used to critically appraise the literature (Coughlan et al, 2013). Thematic analysis and coding were used to combine recurrent ideas that emerged from the literature (Clarke and Braun, 2014). Three themes were derived from the analysis to guide the review:

• Early identification and screening
• Modifiable risk factors
• Preventive interventions.

Early recognition and screening

The prevention of delirium requires routine screening of patients preoperatively and postoperatively to facilitate early identification of symptoms, aid early diagnosis and implement appropriate interventions to reduce its severity. Numerous instruments have been developed to assist health professionals to screen patients for delirium. The advantages of using the Confusion Assessment Method–Intensive Care Unit (CAM-ICU) in reducing POD have been examined by Zanobini et al (2017). In a quasi-experimental study, the researchers used a non-equivalent control group pre-test/post-test design to measure the incidence of POD in intensive care patients before and after the introduction of CAM-ICU. Participants undergoing elective and emergency cardiac surgery were the target population.

In group 1 (n=206) patients were diagnosed with delirium after clinical evaluation from a specialist, and in group 2 (n=153) delirium was assessed by nurses using the CAM-ICU. The CAM-ICU, which was adapted from the Confusion Assessment Method (CAM) developed by Inouye et al (1990), detects delirium alongside the Richmond Agitation Sedation Scale (RASS), which determines whether patients are too sedated to participate during the CAM-ICU assessment. The results showed POD incidence was significantly lower in group 2 compared with group 1 (11.1% versus 41.5%). The reduced prevalence of POD may be due to several factors, including early recognition and treatment of delirium symptoms, pre-treatment differences between the intervention and comparison groups or the fact that clinical evaluation overestimates and the CAM-ICU underestimates the diagnosis of delirium (Zanobini et al, 2017).

Neufeld et al (2013) studied the sensitivity and specificity of the CAM-ICU and the Nursing Delirium Screening Checklist (NuDESC) against a standard neuropsychiatric examination. Although these screening tools do not differ greatly, the NuDESC is designed in a way that enables nurses to assess patients during routine care. Ninety-one participants over the age of 70 years were included. Both the CAM-ICU and NuDESC demonstrated high specificity but insufficient sensitivity for the routine screening of delirium compared with a psychiatric examination in patients admitted to the post-anaesthesia care unit and general postoperative wards. Neufeld et al (2013) recommended a neuropsychiatric examination for the screening of POD; however, psychiatric assessment might not be feasible in some settings, especially when symptoms of delirium are more likely to occur in the evenings, when psychiatric assessments are not easily available.

The NEECHAM Confusion Scale can be used to identify patients considered to be in the high-risk group in which POD frequently develops (Hattori et al, 2009). It assesses patients under three categories:

• Cognition
• Behaviour
• Physiological status based on a scoring system.

The full score is 30 points, and severity increases as the score decreases. Patients with a score of 20 points or less were considered to have severe delirium. Hattori et al (2009) used this instrument to screen 160 participants over the age of 75 years for preoperative cognitive and functional status and surgery-related stress. Patients with dementia and behavioural abnormalities were excluded from the analysis to make the scale testing more reliable. The findings revealed that the preoperative NEECHAM score was accurate in identifying high-risk patients. The incidence of POD was 76% in patients showing a preoperative NEECHAM score of 27 points or less. Thus, the NEECHAM tool may be useful for evaluating the risk of POD to facilitate early prevention strategies.

A prospective observational study (Culley et al, 2017) used the Mini-Cog test, a validated screening tool with high sensitivity and specificity, to detect cognitive impairment in 211 patients aged 65 years and older with no previous diagnosis of dementia. The findings showed that patients with preoperative cognitive impairment, compared with those without, were more likely to develop delirium (Culley et al, 2017). In comparison, Brooks et al (2014) found that only one of 12 patients who tested positive for cognitive impairment using the Mini-Cog assessment had developed POD, whereas delirium occurred in 35 of 211 patients in Culley et al's (2017) study. The difference in results could be because Culley et al's (2017) study only included orthopaedic surgical patients, who have shown increased rates of delirium in contrast with patients undergoing elective surgeries (Bruce et al, 2007).

Modifiable risk factors

Although the aetiology of POD is not well understood, the risk factors for developing it have been well researched. A multifactorial model of the precipitating factors for predicting delirium has been well validated and remains widely accepted (Inouye and Charpentier, 1996). The strongest risk factors are identified as patients being aged over 65 years, followed by cognitive impairment, illness severity, hip fracture and sensorial impairment (National Institute for Health and Care Excellence (NICE), 2010; American Geriatrics Society (AGS), 2015). More research is emerging in this field to identify other risk factors affecting the prevalence of POD, which include potentially modifiable precipitating risk factors, such as malnutrition or pain or anticholinergic burden, which can be influenced by appropriate perioperative strategies (Whitlock et al, 2011; Pasina et al, 2019).

A large observational cohort study (n=1002) was conducted by Radtke et al (2010) to assess for risk factors for POD in the early postoperative period using the NuDESC. The findings showed that preoperative fluid fasting, and the choice of intraoperative analgesic are risk factors for postoperative delirium. Similarly, in a more recent descriptive correlational study by Denny and Lindseth (2017), relationships between subsyndromal delirium (SSD) and preoperative risk factors were examined in older patients following orthopaedic surgery using the CAM method. SSD was defined by Denny and Lindseth (2017) as the presence of one or two core symptoms of delirium without the full range of symptoms, therefore it does not meet the diagnostic criteria. Although both studies measured delirium using different tools, both the CAM and NuDESC were considered valid measurement tools (AGS, 2015). Both studies observed that patients with longer fluid fasting times (over 6 hours) had an increased risk of developing POD.

In relation to anticholinergic burden as a risk factor, Pasina et al's (2019) study assessed the association between delirium and anticholinergic load according to the hypothesis that the cumulative anticholinergic burden increases the risk of delirium. This was a retrospective, cross-sectional study with a sample of older patients (n=477). Pasina et al (2019) concluded that anticholinergic drugs may influence the development of delirium due to the cumulative effect of multiple medications. However, this effect was no longer evident in multivariable logistic regression analysis, after adjustment for dementia and malnutrition.

These findings are supported by Mazzola et al (2017) who conducted a prospective, observational cohort study (n=415) to assess the influence of malnutrition on the prevalence of POD using the Mini Nutritional Assessment–Short Form (MNA-SF). The researchers explained that the MNA-SF is a well-established and validated tool to screen for malnutrition in the hospital setting. Total MNA-SF scores range from 0 to 14, with a score of 7 or less classifying malnourishment. Scores for the MNA-SF are calculated based on information reported by the participant or their proxies, which increases the potential for recall or memory bias (Polit and Beck, 2014). Delirium was assessed by physicians using the CAM in the morning and afternoon but, because of the fluctuating nature of delirium, nurses would also report any changes in the patient at other times as they were observed (Denny and Lindseth, 2017). This helped reduce any underestimation in the findings. Mazzola et al (2017) discovered that participants identified as being malnourished were three times more likely than those with normal nutritional status to develop POD, and those at risk of becoming malnourished were 2.5 times as likely.

Studies on postoperative pain and incidence of POD revealed that higher postoperative pain levels were associated with increased risk of developing delirium in older patients (Leung et al, 2013; Denny and Such, 2018). Leung et al's (2013) prospective cohort study discovered that delirium occurred more frequently among those patients who had a combination of high preoperative risk for delirium, high postoperative pain and were administered high doses of opioid. Denny and Such's (2018) correlational study also showed a relationship between postoperative pain and SSD, while accounting for predisposing risk factors, but they reported that opioid intake was not significantly associated with SSD. However, this study had a smaller sample size compared with Leung et al (2013) (49 participants versus 581 participants), so its findings should be interpreted cautiously.

Preventive interventions

In hospitalised patients, 30–40% cases of delirium are thought to be preventable (Zhang et al, 2013). Delirium prevention is regarded as an indicator of high-quality care and a cost-effective strategy (Inouye, 2006; NICE, 2010). Given that delirium is a complex problem resulting from multiple factors, it follows that the approach to prevention needs to be multifactorial (Wass et al, 2008). Three experimental studies explored the impact of multifactorial interventions in older hip-fracture patients. Marcantonio et al's (2001) randomised controlled trial (RCT) introduced proactive geriatric assessment to provide targeted recommendations to improve patient conditions after surgery. The recommendations described in this study were similar to the interventions implemented by Björkelund et al (2010) in a study that evaluated the efficacy of early interventions, such as supplemental oxygen, increased fluid and nutrition intake, reducing polypharmacy, daily delirium screening, adequate pain relief and regular monitoring of vital physiological parameters in reducing POD.

Freter et al (2017) evaluated the effectiveness of delirium-friendly pre-printed orders (PPOs) with routine postoperative orders. These included options and doses for night sedation, analgesia, nausea with attention to catheter removal and bowel movements. Marcantonio et al (2001) showed that proactive geriatric assessment reduced the incidence of POD by over one third, but severe cases of delirium were reduced by one half. This finding is supported by the findings of a systematic review and meta-analysis by Moyce et al (2014), which suggested that perioperative geriatric assessment with multicomponent interventions are effective in decreasing the incidence of POD.

A nurse-led intervention in the ICU was successful in reducing delirium in a quasi-experimental study (Guo and Fan, 2016). This study emphasised key factors such as education of nursing staff, providing emotional and social support to patients and families, maintaining a safe environment, and improving sleep quality in patients after surgery. It was noted that patients who received these interventions had lower scores on the Delirium Detection Scale, which may indicate that patients in the intervention group were also recipients of higher quality care compared with those in the usual care group.

An intervention study by Kratz et al (2015) and an RTC by Chen et al (2017) evaluated the Hospital Elder Life Program (HELP), with Chen et al using the modified HELP (mHELP) This is an innovative strategy to provide care for older patients that uses tested delirium prevention interventions to improve the overall quality of care. HELP was implemented by specialist nurses in both studies perioperatively and included the following:

• Maintaining orientation to surroundings and providing visual and hearing aids for patients with sensory impairments
• Meeting needs for nutrition, fluids and sleep
• Promoting mobility and involvement of relatives.

In these studies (Kratz et al, 2015; Chen et al, 2017), patients in the intervention group who were enrolled in HELP had a lower risk of POD, with Chen et al (2017) showing a reduction in delirium risk of 56%. Furthermore, Kratz et al (2015) suggested that a delirium liaison nurse could contribute to reducing the incidence of POD.

An RCT conduced by Lee et al (2013) evaluated whether a perioperative psycho-educational intervention could reduce POD by reducing surgery-related anxiety. Participants in the study were divided into a control group (n=46) who received standard care, and an intervention group (n=49) who received preoperative and postoperative reassurance and education regarding care in the ICU setting, possibility of complications including delirium, explanation of procedures and future treatment plans. The incidence of delirium within the first week postoperatively was significantly lower in the intervention group compared with the controls.

In another study, the impact of multifaceted preoperative patient education on anxiety, knowledge and POD was evaluated by Chevillon et al (2015). The experimental group (n=63) received information from ICU nurses trained to provide specific individualised education, while the control group (n=66) received unstructured, preoperative education from all members of the multidisciplinary team. Findings showed that the two groups did not differ significantly in anxiety, incidence of delirium or ICU days. Other benefits of individualised patient education were significantly more knowledge about postoperative care (P<0.001) and fewer days of mechanical ventilation (P=0.04) in the intervention group than the control group.

The use of antipsychotic prophylaxis to reduce POD is supported by a study undertaken by Khan et al (2018). A total of 135 patients undergoing thoracic surgery were enrolled in a randomised double-blind placebo-controlled trial in the ICU, which included a subgroup of patients undergoing oesophagectomy (n=84). Patients in the experimental group received low doses of intravenous haloperidol (0.5 mg three times daily) postoperatively and those in the control group were administered a placebo identical in route, appearance and volume. The researchers discovered that low doses of haloperidol did not reduce the incidence or severity of delirium in patients undergoing thoracic surgery, but found that there was a 17% reduction in delirium in oesophagectomy patients.

Teslyar et al (2013) and Fok et al (2015) explored the role of preoperative and postoperative antipsychotic medications in decreasing POD. Both systematic reviews with meta-analysis focused on randomised placebo-controlled trials and followed a robust search strategy and study screening. Teslyar et al (2013) analysed five studies (1491 participants), while Fok et al (2015) examined six studies (1710 participants). Antipsychotics were shown to reduce the incidence of delirium in several surgical settings in both reviews. The most common antipsychotic medications used were haloperidol, risperidone and olanzapine, but neither of the studies could identify which antipsychotic might offer optimal prophylaxis against delirium. However, a major limitation of prophylactic antipsychotics use are associated adverse effects such as sedation and constipation, especially in the older population.

Conclusion

The aim of this review was to explore prevention strategies for POD in older adults. This review has shown that POD was a common complication in older patients globally. From the literature, it was evident that early recognition and screening, identification of modifiable risk factors and specific non-pharmacological interventions were the most widely used strategies to prevent POD.

Delirium is not easily detected, but nurses have a crucial role in observing patients, identifying symptoms and communicating the symptoms of delirium to doctors. Using reliable and validated instruments, such as the CAM-ICU or NuDESC, may help to support the ability of nurses to easily recognise symptoms, but also initiate early treatment to reduce the incidence and impact of POD. The introduction of the NEECHAM scale into standard preoperative nursing care could be beneficial for screening high-risk patients and help to provide early, targeted interventions. Patients over 65 years of age should also be routinely screened preoperatively using Mini-Cog assessments, because cognitive impairment before surgery appears to be a significant predictor of POD.

Risk assessment of predisposing and precipitating factors should start preoperatively to initiate timely interventions. The findings of the review indicate that preoperative fasting times should be reduced as part of delirium prevention; however, more interventional studies are needed to confirm any conclusions regarding shorter fluid fasting times and a reduction in POD. Improving the assessment of malnourishment and overall nutrition in hospitalised patients should be prioritised, because the risk of developing POD is three times as high if patients are malnourished. Patients with high postoperative pain levels are more likely to develop POD, therefore, nurses should aim to provide effective pain relief to older adults during the postoperative period, while regularly observing patients for adverse effects of medications, especially opioids.

Due to the negative outcomes and increased healthcare costs associated with POD, proactive geriatric assessment and multifactorial interventions focused on modifying precipitating risk factors are important. Little research has focused on nursing interventions for delirium prevention, but strategies such as HELP have proved to be beneficial in the surgical setting. The findings from the studies indicate that antipsychotics need to be used with caution and only when the benefits outweigh the risks. The standardisation of screening for POD preoperatively would help increase detection rates and help initiate early intervention to reduce complications after surgery.

In addition, healthcare providers should facilitate the inclusion of family members in the postoperative treatment of older patients, because those close to the patient may know the patient's baseline cognitive status and will be able to recognise early signs of delirium. Education and training should be provided to nurses involved in the postoperative care of older patients. Nurses need to improve their understanding of the relationship between older age and surgical complications, specifically neurological complications, and so future studies should aim to develop nursing interventions for delirium prevention.

Key Points

• Early prevention of postoperative delirium (POD) is the most effective management strategy to reduce incidence
• Nurses should prioritise systematic screening of older at-risk patients preoperatively and postoperatively using validated instruments to detect delirium early and reduce its impact
• Healthcare professionals should aim to minimise precipitating factors such as prolonged fasting times, malnutrition and postoperative pain, making patients less susceptible to POD
• Multifactorial interventions and proactive geriatric assessment focusing on non-pharmacological and pharmacological strategies have shown to be most beneficial in reducing delirium following hip-fracture surgery

CPD reflective questions

• What policies or guidelines do you use in practice for the detection, prevention and management of postoperative delirium (POD)?
• What support to you provide in educating patients and families in identifying and understanding the symptoms of POD preoperatively in order to reduce the impact of POD?
• What precipitating risk factors that increase the risk of developing POD in older patients can be modified or removed in your area of practice?