Risk factors and prevention for postoperative delirium after orthopedic surgery.

BACKGROUND: Postoperative delirium significantly delays the recovery of patients. This study sought to explore the risk factors and to prevent postoperative delirium after orthopedic surgery.
MATERIALS AND METHODS: In the part of retrospective study, 456 cases over 70-year-old that underwent one of three types of orthopedic surgery were included and were defined as the retrospective group. The risk factors of postoperative delirium were analyzed by logistic regression. In the part of prospective study, 86 cases of the same age and the same surgery were included and were defined as the prospective group. Positive interventions were used by shortening the fasting time and increasing the perioperative rehydration. The differences of postoperative delirium incidence between the retrospective group and the prospective group were analyzed.
RESULTS: Compared with patients with normal postoperative electrolytes, postoperative creatinine <68.20 μmol/L, and fluid infusion during fasting >119.66 mL/h, postoperative electrolyte disorders (odds ratio [OR]: 2.864; 95% confidence interval [CI]: 1.374, 5.970), postoperative creatinine ≥68.20 μmol/L (OR: 2.660; 95% CI: 1.328, 5.328), and fluid infusion during fasting ≤119.66 mL/h (OR: 2.372; 95% CI: 1.197, 4.704) were the risk factors for postoperative delirium. After positive intervention, the postoperative delirium incidence of the prospective group was 5.8% (5/86), and it was lower than 18.4% (84/456) of the retrospective group (P < 0.05).
CONCLUSIONS: Elevated postoperative creatinine, postoperative electrolyte disorders, and lower fluid infusion during fasting were three risk factors for postoperative delirium. By shortening the fasting time and increasing the perioperative rehydration, the incidence of postoperative delirium could be reduced. Copyright:

INTRODUCTION

Postoperative delirium is an acute cognitive impairment that exhibits clinical features such as nonspecific and acute disturbance of attention, consciousness, sleep-wake cycles, and cognitive ability.[1] Numerous surgical cases of various types have clearly shown that patients often develop postoperative delirium.[23] Postoperative delirium is common in orthopedic wards after hip replacement[4] fracture fixation,[5] knee replacement,[6] and lumbar decompression and lumbar fusion surgical procedures.[7] Postoperative delirium is generally a negative factor in rehabilitation, and it increases medical costs and delays the discharge of patients from the hospital. The delirium can also lead to many serious complications such as bedsores, fall-related fractures, and even death.[8]

In many risk prediction models for postoperative delirium, age was the most often used predictor.[9] Our previous research has also found that aging and postoperative electrolyte disorders were risk factors for postoperative delirium in patients undergoing orthopedic surgeries.[10] However, not all elderly patients will experience postoperative delirium. What is the cause of postoperative electrolyte disorders? Is there a way to prevent and reduce postoperative delirium?

We retrospectively enrolled elderly patients of the same age undergoing orthopedic surgery to identify possible causes for postoperative delirium. Moreover, in the part of prospective study, positive interventions were used to test whether postoperative delirium could be effectively prevented.

MATERIALS AND METHODS

Ethical considerations

All protocols were approved by the Internal Review Board of Dongyang People's Hospital.

Study design and participants

In our ward, for patients older than 60 years of age, we routinely screen delirium by the Confusion Assessment Method when admitted to hospital. Based on a review of medical charts, each patient with delirium had a record of consultation with a senior psychiatrist. The psychiatrists made an independent diagnosis based on the Diagnostic and Statistical Manual of Mental Disorders, 4th edition, Text Revision[11] and proposed a treatment plan for each delirium case.

Between January 2015 and November 2018, as the part of retrospective study, all patient information was obtained from medical chart review. Younger patients are significantly less likely to develop postoperative delirium than older patients.[9] To rule out the effects of age, we selected surgical patients over 70 years of age as the subjects. Under general anesthesia, 540 individuals who underwent one of the three types of orthopedic surgeries namely internal fixation for intertrochanteric fracture (IFIF), femoral head replacement (FHR), and total hip arthroplasty (THA) were enrolled. Among these cases, 84 were excluded: three cases of simultaneous bilateral surgery; 35 cases with mental illness, concurrent prescription of psychiatric drugs or preoperative cognitive impairment such as dementia and delirium (6 cases of preoperative delirium); 15 cases with severe vision or hearing disorders who could not complete the cognitive function tests; 16 cases with cerebral infarction or cerebral hemorrhage; and 15 cases who were lost to follow-up or for whom the follow-up data were incomplete. Thus, a total of 456 cases, including 222 cases of IFIF, 69 cases of FHR, and 165 cases of THA were included in this retrospective study and were defined as the retrospective group.

Among the 456 cases undergoing orthopedic surgery, 84 cases of postoperative delirium were defined as the delirium group, and 372 cases without postoperative delirium were defined as the control group [Figure 1]. Using multivariate logistic regression, we statistically analyzed the risk factors of postoperative delirium.

Figure 1

The flow diagram of the retrospective study part

On the base of risk factors for postoperative delirium, we speculated that electrolyte disorders and temporary renal function decline caused by inadequate infusion were important triggers for postoperative delirium. Between December 2018 and July 2019, it was the part of prospective study. By shortening the fasting time and increasing the perioperative rehydration, we hoped to reduce the incidence of postoperative delirium for patients over 70-year-old who would be performed one of the above three procedures. All procedures were performed under general anesthesia. The reduction of fasting time was mainly achieved by preoperative oral rehydration. While monitoring the patients’ heart function, the perioperative rehydration has been appropriately increased. Finally, a total of 86 cases, including 41 cases of IFIF, 15 cases of FHR, and 30 cases of THA were included in this prospective study, and were defined as the prospective group. The differences in incidence of postoperative delirium between the retrospective group and the prospective group have been analyzed.

Data collection

Age, gender, and perioperative factors were collected through medical records; these factors included preoperative hospitalization, type of surgery, preoperative albumin, preoperative hemoglobin, preoperative creatinine, preoperative urea nitrogen, preoperative alanine aminotransferase, preoperative total bilirubin, postoperative intensive care unit (ICU) care, postoperative creatinine, postoperative urea nitrogen, postoperative alanine aminotransferase, postoperative total bilirubin, postoperative electrolyte disorders (potassium, sodium, chlorine, and calcium), fasting time, and fluid infusion per hour during fasting. Electrolyte disorder was defined as values outside the range, potassium (3.5–5.5 mmol/L), sodium (135–145 mmol/L), chlorine (95–105 mmol/L), and/or calcium (2.25–2.58 mmol/L). The levels of albumin and hemoglobin were measured preoperatively. The serum concentrations of creatinine, urea nitrogen, alanine aminotransferase, and total bilirubin were determined preoperatively and 1 day postoperatively. Potassium, sodium, calcium, and chlorine were measured 1 day postoperatively. Fluid infusion per hour during fasting was calculated by total infusion volume (including blood transfusion) and fasting time.

Statistical analysis

Continuous variables were presented as the mean ± standard derivation, and a Student's t-test or t’-test (unequal variances) was used to determine whether there was a significant difference between the two groups [Tables 1 and 2]. To examine the effect of demographic and perioperative factors on the incidence of postoperative delirium, univariate logistic regression was employed first, followed by multivariate logistic regression analyses with the forward stepwise (conditional) method in SPSS [Tables 3 and 4]. Chi-square test was applied to analyze the difference of incidence or ratio between the retrospective group and the prospective group [Table 2]. All statistics were performed with SPSS version 18.0 (SPSS, Inc., Chicago, IL, USA), and the level of significance was set at P < 0.05.

Table 1

Statistical analysis of continuous variables between the delirium group and the control group

	Delirium group	Control group	t a	P
Cases	84	372	-	-
Age (years)	80.18±6.44	79.31±8.01	0.99	0.32
Preoperative hospitalization (d)	5.15±2.63	5.31± 3.93	0.30	0.76
Preoperative albumin (g/L)	37.63±3.59	37.44±3.99	0.31	0.76
Preoperative hemoglobin (g/L)	114.49±14.63	115.00±18.10	0.19	0.85
Preoperative creatinine (μmol/L)	75.55±28.22	69.41±26.56	1.34	0.16
Preoperative urea nitrogen (mmol/L)	6.70±2.52	7.20±2.31	1.30	0.19
Preoperative alanine aminotransferase (U/L)	13.41±4.96	16.85±10.80	2.50	0.01
Preoperative total bilirubin (μmol/L)	15.12±7.94	16.39±7.95	1.00	0.32
Postoperative creatinine (μmol/L)	76.68±30.12	66.29±24.52	2.37	0.02
Postoperative urea nitrogen (mmol/L)	8.81±3.52	7.71±3.45	1.96	0.051
Postoperative alanine aminotransferase (U/L)	15.26±9.72	15.13±9.07	0.09	0.93
Postoperative total bilirubin (μmol/L)	17.53±7.75	18.55±9.02	0.75	0.45
Fasting time (h)	19.11±2.31	19.37±2.42	0.69	0.49
Fluid infusion per hour during fasting (mL/h)	104.38±24.64	123.11±21.45	6.99	<0.001

aA Student’s t-test or t′-test (unequal variances) was used to analyze these variables

Table 2

Postoperative delirium incidence of the retrospective group and the prospective group

	Retrospective group	Prospective group	P
Cases	456	86	-
Postoperative delirium (yes/no)	84/372	5/81	<0.001
Type of surgery (cases)			0.86
IFIF	222	41
FHR	69	15
THA	165	30
Age (years)	79.47±7.52	80.01±8.05	0.56
Gender (male/female)	162/294	29/57	0.75
Postoperative creatinine (μmol/L)	68.20±27.16	66.37±20.96	0.55
Postoperative electrolyte disorders (yes/no)	269/187	35/51	<0.001
Fasting time (h)	19.32±2.47	12.12±3.04	<0.001
Fluid infusion per hour during fasting (mL/h)	119.66±22.38	144.57±23.39	<0.001

IFIF – Internal fixation for intertrochanteric fracture; FHR – Femoral head replacement; THA – Total hip arthroplasty

Table 3

Univariate logistic regression analysis for postoperative delirium

Characteristics	B a	SE	Wald	P	OR	95% CI
Gender
Male	-	-	-	-	1.000	-
Female	−0.484	0.344	1.987	0.159	0.616	0.314-1.208
Type of surgery			1.892	0.595
IFIF	-	-	-	-	1.000	-
FHR	0.269	0.425	0.402	0.526	1.309	0.570-3.009
THA	−0.185	0.410	0.203	0.652	0.831	0.372-1.857
Postoperative ICU care
No	-	-	-	-	1.000	-
Yes	0.460	0.327	1.981	0.159	1.585	0.835-3.009
Postoperative creatinine (μmol/L)
<68.20	-	-	-	-	1.000	-
≥68.20	0.982	0.337	8.491	0.004	2.669	1.379-5.166
Postoperative urea nitrogen (mmol/L)
≤7.14	-	-	-	-	1.000	-
>7.14	1.027	0.339	9.160	0.002	2.794	1.436-5.434
Postoperative electrolyte disorders
No	-	-	-	-	1.000	-
Yes	0.899	0.350	6.601	0.010	2.458	1.238-4.880
Potassium
Normal	-	-	-	-	1.000	-
Disorder	0.590	0.554	1.138	0.286	1.805	0.610-5.340
Sodium
Normal	-	-	-	-	1.000	-
Disorder	0.306	0.352	0.756	0.385	1.358	0.681-2.707
Chlorine
Normal	-	-	-	-	1.000	-
Disorder	0.325	0.431	0.568	0.451	1.384	0.594-3.223
Calcium
Normal	-	-	-	-	1.000	-
Disorder	0.706	0.327	4.651	0.031	2.026	1.067-3.850
Fluid infusion per hour during fasting (ml/h)
≤119.66	-	-	-	-	1.000	-
>119.66	−0.832	0.328	6.445	0.011	0.435	0.229-0.827

aRegression coefficient. SE – Standard error; OR – Odds ratio; CI – Confidence interval; IFIF – Internal fixation for intertrochanteric fracture; FHR – Femoral head replacement; THA – Total hip arthroplasty; ICU – Intensive care unit

Table 4

Multivariate logistic regression analysis for postoperative delirium

	B a	SE	Wald	P	OR	95% CI
Postoperative creatinine (μmol/L)
<68.20	-	-	-	-	1.000	-
≥68.20	0.978	0.354	7.617	0.006	2.660	1.328-5.328
Postoperative electrolyte disorders
No	-	-	-	-	1.000	-
Yes	1.052	0.375	7.882	0.005	2.864	1.374-5.970
Fluid infusion per hour during fasting (mL/h)
≤119.66	0.864	0.349	6.120	0.013	2.372	1.197-4.704
>119.66	-	-	-	-	1.000	-

aRegression coefficient. SE – Standard error; OR – Odds ratio; CI – Confidence interval

RESULTS

Incidence in retrospective study and onset of delirium

Among the 456 individuals over 70-year-old who underwent one of the three types of orthopedic surgeries, the incidence of postoperative delirium was 18.4% (84/456). Through treatment of haloperidol, rehydration, and regaining electrolyte balance, most delirium patients showed improvement within 1–2 weeks; however, two patients still left cognitive dysfunction at the last follow-up. No mortality during hospitalization occurred because of postoperative delirium.

Student's t-test or t’-test for the continuous variables

Patients with postoperative delirium had lower levels of preoperative alanine aminotransferase, fluid infusion per hour during fasting (59 cases lower than the average), and higher levels of postoperative creatinine (48 cases higher than the average) [P < 0.05, Table 1]. Postoperative creatinine and fluid infusion per hour during fasting were then transformed into dummy variables [Table 3] and were used in the univariate logistic regression analysis. Because the P value was close to 0.05, postoperative urea nitrogen (57 cases in the delirium group higher than the average) was also included in the subsequent analysis. The level of preoperative alanine aminotransferase in the control group was higher than that in the delirium group, and it was clear that preoperative alanine aminotransferase was not the cause of postoperative delirium. Therefore, preoperative alanine aminotransferase were excluded. In the end, three important variables were entered into the next univariate logistic regression analysis, which were postoperative creatinine, postoperative urea nitrogen, and fluid infusion per hour during fasting.

Univariate logistic regression analysis

In this univariate logistic regression analysis, postoperative delirium was set as the dependent variable, and 11 factors were set as the independent variables (gender, type of surgery, postoperative ICU care, postoperative creatinine [more or less than the average, 68.20 μmol/L], postoperative urea nitrogen [normal or more than 7.14 mmol/L], postoperative electrolyte disorders, abnormal potassium, sodium, chlorine, calcium levels, and fluid infusion per hour during fasting [more or less than the average, 119.66 mL/h]) [Table 3]. Univariate logistic regression analysis revealed that five factors were associated with postoperative delirium, namely postoperative creatinine, postoperative urea nitrogen, postoperative electrolyte disorders, abnormal calcium level, and fluid infusion per hour during fasting [P < 0.05, Table 3].

Multivariate logistic regression analysis

Subsequently, the five factors associated with postoperative delirium [Table 3] were analyzed statistically using multivariate logistic regression. The analysis identified three risk factors for postoperative delirium in patients undergoing orthopedic surgery: elevated postoperative creatinine, postoperative electrolyte disorders, and lower fluid infusion per hour during fasting [Table 4].

Comparison of postoperative delirium incidence between the retrospective group and the prospective group

By shortening the fasting time and increasing the perioperative rehydration, the postoperative delirium incidence of the prospective group was 5.8% (5/86), and it was lower than 18.4% (84/456) of the retrospective group [P < 0.05, Table 2]. Compared with the retrospective group, the incidence of postoperative electrolyte disorders in the prospective group was lower [P < 0.05, Table 2].

DISCUSSION

Among elderly patients, postoperative delirium is a common phenomenon after various types of surgery. Postoperative delirium leads to recovery delay, increased medical costs and in some cases even serious complications such as fall-related fractures and deaths. In this study, we found that elevated postoperative creatinine level, postoperative electrolyte disorders, and lower fluid infusion per hour during fasting were three main risk factors for postoperative delirium. By shortening the fasting time and increasing the perioperative rehydration, the incidence of postoperative delirium and postoperative electrolyte disorders have been reduced.

Several studies have demonstrated that fluid infusion may correlate with postoperative delirium. Olin et al.[12] reported that delirium patients had significantly greater intraoperative blood loss and intravenous fluid infusion than those without postoperative delirium after major abdominal surgery. Greater blood loss means that greater fluid infusion is necessary. Echigoya's and Kato's study[13] confirmed that permissive hypotension is a risk factor for postoperative delirium. Because hypotension is closely related to inadequate transfusion, fluid infusion, long operation time, and substantial blood loss, we believe that inadequate fluid infusion is a risk factor for postoperative delirium.

Serum creatinine concentration is an important biomarker for monitoring renal function.[1415] Although no concrete evidence exists for an association between renal dysfunction and postoperative delirium, some studies have shown that renal dysfunction is related to psychiatric disorders such as delirium and coma. For example, in an elderly patient with chronic kidney disease and silent ischemic cerebrovascular disease, Lin's and Chew's study showed that episodes of seizure and delirium were associated with intravenous ertapenem and creatinine clearance level.[16] Similarly, in our study, the high level of postoperative creatinine was a risk factor for postoperative delirium. Siew et al. found that acute kidney injury is a risk factor for delirium and coma in a prospective study of ICU patients with shock and/or respiratory failure. Moreover, renal replacement therapy has confirmed the association between kidney injury and both delirium and coma.[15]

Many studies have indicated that electrolyte disorders are associated with some neuropsychiatric disorders such as dementia,[17] Alzheimer's disease,[18] and depression.[19] Cherbuin's study even showed that intake of potassium, calcium, and magnesium can reduce the risk of vascular dementia.[20] The present study confirmed the role of postoperative electrolyte disorders in postoperative delirium. The risk of postoperative delirium in individuals with electrolyte disorders was 2.9 times that of patients with normal electrolyte levels. In the delirium guideline,[21] an electrolyte disorder is a conclusive risk factor for postoperative delirium. An abnormal potassium or sodium level is commonly associated with an imbalance of body fluids, metabolic alkalosis, or microcirculation disorders. Undoubtedly, these conditions can lead to certain psychiatric symptoms including apathy, depression, confusion, drowsiness, fatigue or even coma, which have been verified by some previous studies.[722]

However, the effects of different electrolytes on the occurrence of postoperative delirium remain controversial. Some studies have indicated that an abnormal in sodium or potassium level might be a contributing factor in postoperative delirium,[23] but others have reported opposite results.[24] In our univariate logistic regression analysis, an abnormal calcium level was associated with postoperative delirium, but this could not be identified in a multivariate logistic regression analysis. Among our cases, the most common calcium disorder was hypocalcemia, which was also taken as a risk factor for postoperative delirium in Aldemir et al.'s study.[25]

We speculate that postoperative electrolyte disorders resulting from inadequate fluid infusion and temporary renal dysfunction are important causes of postoperative delirium. Thus, an effective balance of fluid and electrolytes is maybe important for reducing the occurrence of postoperative delirium. This speculation has been partly confirmed in our prospective study. By shortening the fasting time and increasing the perioperative rehydration, the incidence of postoperative delirium and postoperative electrolyte disorders has been reduced.

This study has some primary limitations. First, postoperative delirium is a common phenomenon after orthopedic surgery. The three types of surgery we selected may not represent all types of orthopedic surgery. Second, further study is necessary to identify which specific electrolyte disorders lead to postoperative delirium. Finally, many factors may be related to postoperative delirium, not all related factors were included in this study, such as anesthesia.[26]

CONCLUSIONS

Elevated postoperative creatinine, postoperative electrolyte disorders, and lower fluid infusion per hour during fasting were three main risk factors for postoperative delirium after orthopedic surgery. The incidence of postoperative delirium could be reduced by shortening the fasting time and increasing the perioperative rehydration.

Financial support and sponsorship

This research was supported by the Department of Health of Zhejiang Province, China (2011KYB152).

Conflicts of interest

There are no conflicts of interest.