Sage Journals: Discover world-class research

Abstract

Purpose: Aspiration pneumonia is a critical issue. General anesthesia may suppress the airway’s protective reflex. However, aspiration pneumonia is also observed in patients who undergo hip fracture surgery under spinal anesthesia. The aim of this study was to investigate the relationship between anesthesia methods and aspiration pneumonia as well as the predictive factors of aspiration pneumonia in elderly patients undergoing hip fracture surgery. Methods: The medical records of 19,809 patients aged ≥60 years who underwent hip fracture surgery under general or spinal anesthesia were reviewed. After propensity score matching, the anesthesia methods affecting the occurrences of aspiration pneumonia and other complications were investigated via logistic regression and instrumental variable analyses. Predictive factors of aspiration pneumonia were also investigated in all subjects using a multivariable logistic regression analysis. Results: Among the 11,673 general anesthesia patients and 8136 spinal anesthesia patients, aspiration pneumonia occurred in 356 patients (1.8%). Post-propensity score matching the incidences of aspiration pneumonia with general and spinal anesthesia were 1.8% and 1.5%, respectively (p = 0.158); other pulmonary complications were 1.5% and 1.5%, respectively (p = 0.893); and the mortality rates were 1.4% and 1.2%, respectively (p = 0.219). The predictive factors of aspiration pneumonia were advanced age, male sex, lean body, cerebrovascular disease, dementia, and dependency for activities of daily living (eating). Conclusion: Spinal and general anesthesia showed similar incidences of aspiration pneumonia in elderly hip fracture surgery. Regardless of the anesthesia method, great care should be taken, especially in elderly patients with the identified predictive factors.

Keywords

aspiration pneumonia elderly general anesthesia hip fracture surgery spinal anesthesia

Introduction

Aspiration pneumonia occurs in approximately 1% of surgical patients,¹ and perioperative aspiration pneumonia is associated with significantly high mortality.^1–3 The frequency of aspiration pneumonia increases with age.⁴ Similarly, the frequency of osteoporotic fractures also increases with age. The number of hip fractures was about 1.6 million worldwide in 2000, and it is estimated to reach 4.5 million by 2050.^5,6

General anesthetics and/or airway interventions inhibit laryngeal defensive reflexes.^7,8 Since upper airway reflex sensitivity decreases with increasing age,^9,10 general anesthesia may increase the risk of aspiration pneumonia in elderly patients. However, current general anesthetics have a rapid elimination time.^11–13 Since most patients undergo hip fracture surgery after an appropriate fasting period, aspiration of gastric contents during anesthesia induction is unlikely. Furthermore, Studer et al.³ reported that postoperative aspiration pneumonia occurs at 7 ± 10 days after surgery. It is still unclear whether general anesthesia is associated with postoperative aspiration pneumonia in elderly hip fracture surgery patients. Therefore, the aim of this study was to investigate whether anesthesia methods are related to the occurrence of aspiration pneumonia in elderly patients undergoing hip surgery. We also investigated the predictive factors of postoperative aspiration pneumonia.

Methods

The study was approved by the ethical review boards of our hospital (2019-1; principal investigator, T.F.; date of approval, January 2019) and the National Hospital Organization (NHO). Following the Japanese ethical guidelines for human medical research, which are based on the Declaration of Helsinki, the study protocol was open to the general public via websites in order to obtain patient objections. To protect patient privacy, all personal identification data were encrypted in a security room of the NHO data bank. The requirement for written informed consent was waived by the ethical review boards.

Data sources

A retrospective cohort study was conducted using data from an administrative database. The research period was from 1 April 2010, to 31 March 2019. Data from 81 NHO hospitals were used. The data were obtained from the Japanese Diagnosis Procedure Combination (DPC) administrative claims database. The DPC database is a diagnosis-dominant, case-mix system administered by Japan’s Ministry of Health, Labour and Welfare, and it is linked with a lump-sum payment system. The data include the primary diagnosis, the diagnosis on admission, the most and second-most resource-consuming diagnoses, 10 comorbid diagnoses on admission, and 10 complications during hospitalization. Hospital scale data were obtained from the NHO’s resident recruitment guidebooks and each hospital’s website.

Selection of patients and variables

The inclusion criteria were as follows: (1) patients who underwent surgery during the research period for a hip fracture (transcervical, intertrochanteric, or subtrochanteric femoral fracture) within 4 days of hospitalization, including the day of hospitalization; (2) patients who underwent total arthroplasty, hemiarthroplasty, or another typical hip fracture procedure (plate/screw or intramedullary implant); and (3) patients who underwent either general anesthesia or spinal anesthesia. The general anesthesia group included patients who received a combination of general anesthesia and an epidural block. Patients who received unknown anesthesia, mixed spinal and general anesthesia, or other anesthesia types were excluded. We also excluded younger patients (<60 years old).

The data from the DPC database included age, sex, height, weight, smoking status, fracture type, surgery date, surgical procedure, drug treatment (sedatives), comorbidities on admission, diseases occurring after admission, and functional status, which was based on the patient’s ability to perform tasks related to activities of daily living (ADLs) before admission. Body mass index (BMI) was calculated from height and weight. A BMI between 25 kg/m² and 29.9 kg/m² is considered overweight, while a BMI ≥30 kg/m² is considered obese.¹⁴ However, many Japanese people are relatively thin. In this study, only 1% of subjects had a BMI ≥30 kg/m². Therefore, BMI was divided into three groups: < 18.5 kg/m², 18.5–24.9 kg/m², and ≥25 kg/m². All diagnoses were defined using the codes of the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10). Surgical procedures were defined using Japanese procedure codes. Surgical cost, including technical fees and material costs, was selected for compensate classification of the surgical procedures. Functional status includes the following parameters: (1) eating, (2) transferring, (3) grooming, (4) toileting, (5) bathing, (6) walking on a flat floor, (7) using stairs, (8) dressing, (9) defecating, and (10) micturating.¹⁵ Five parameters (eating, grooming, toileting, bathing, and walking) were selected from these 10 ADLs as indicators of physical weakness.¹⁶ The age-adjusted Charlson Comorbidity Index (CCI) score was calculated for each patient based on Quan et al.’s coding algorithms.^17,18

The first outcome variable was the occurrence of aspiration pneumonia, identified as J69.0 in the ICD-10. The definition of J69.0 is “pneumonia due to inhalation of food and vomit.” This definition does not take into account the severity of aspiration pneumonia. The second outcome variables were the occurrences of pneumonia (J13–18), bronchitis (J20–22), atelectasis (J981), pneumothorax (J93), pulmonary edema (J81), pulmonary embolism (I26), and death in the hospital. We defined these variables using the ICD-10 code recordings of the diagnosis categories of the complications during hospitalization.

Propensity score matching

A propensity score-matched analysis without replacement was used to reduce the selection bias and potential baseline differences between the general and spinal anesthesia groups. Scores were calculated using a binary logistic regression analysis, with anesthesia type as the dependent variable and 26 preoperative factors as the predictor variables. The independent variables included age, sex, BMI, age-adjusted CCI score, smoking status, delirium, fracture type, surgery type, surgical cost, days from hospitalization to surgery, sedative administration, hospital scale (number of beds), calendar date (year), pre-hospitalization ADL parameters (eating, grooming, toileting, bathing, and walking), and some comorbidities that are closely related to the choice of anesthesia method: atrial fibrillation/atrial flutter (I48), cerebrovascular diseases (I60–69), coagulopathy (D65–69), mental disorders (F00–99), neurologic diseases (G00–99), respiratory diseases (J00–99), thrombophlebitis (I80), and vertebral diseases (M40–54). The C-statistic was used to evaluate the discriminatory ability of the model. The caliper was set at 0.2 times the standard deviation of the estimated propensity score. Standardized differences were used to compare the characteristics between the two groups before and after matching.

Statistical analysis

The preoperative variables were compared between the general and spinal anesthesia groups using the chi-square test. ADL independence was compared between patients with and without aspiration pneumonia using the chi-square test. The occurrences of aspiration pneumonia, other pulmonary complications, and mortality were assessed via logistic regression analysis after propensity score matching. Odds ratios and 95% confidence intervals were then estimated.

Hidden biases caused by unmeasured confounders, such as the hospital’s characteristics, may not have been removed with the propensity score-matched analysis. Therefore, we performed an instrumental variable analysis and a sensitivity analysis after propensity score matching.¹⁹ Based on the proportions of general anesthesia and spinal anesthesia performed at the NHO hospitals between 1 April 2019, and 31 March 2020, we classified each hospital into one of two groups—one with a high proportion of general anesthesia and the other with a low proportion of general anesthesia—and used this classification as an instrumental variable. A two-stage residual inclusion model was used to estimate the effect of each anesthesia method on the outcomes.¹⁹

To investigate the predictive factors of aspiration pneumonia, a multivariable logistic regression analysis was performed between aspiration pneumonia-positive and -negative patients. In addition to anesthesia methods, the following factors were selected for the model: age, sex, BMI, cerebrovascular disease, Parkinson disease/syndrome, dementia, and initial ADL ability (eating). These variables were previously suggested to have an effect on postoperative pulmonary complications.^1,2,20–22

A p-value < 0.05 was considered significant for all statistical tests. Statistical analyses were performed using the ivtools package²³ in R software version 4.0.4 (R Core Team, Vienna, Austria) for the instrumental variable analysis and SPSS software version 26 (IBM, Armonk, New York, USA) for all other analyses.

Results

Figure 1 shows the patient selection process. In the DPC database, a total of 39,800 patients who underwent hip fracture surgery during the observation period (1 April 2010 to 31 March 2019) were identified. Among the identified patients, a total of 19,991 were excluded because other anesthesia methods were used, there were multiple surgeries, the patient was under 60 years of age, or surgery was delayed by more than 4 days. There were 19,809 patients who met the inclusion criteria (11,673 general anesthesia patients and 8136 spinal anesthesia patients). Propensity score matching yielded 7519 pairs, and the C-statistic for the logistic regression was 0.651 (95% confidence interval: 0.643–0.658).

Figure 1.

Flow diagram of patient inclusion and exclusion for analysis of postoperative aspiration pneumonia.

For the 19,809 patients, the morbidity of aspiration pneumonia was 1.8% (356 patients). For patients in their 60s, 70s, 80s, and ≥90s, the morbidities of aspiration pneumonia were 0.7%, 1.3%, 1.8%, and 2.4%, respectively. For aspiration pneumonia-positive and -negative patients, the mortalities were 19.1% and 1.0%, respectively (p < 0.001).

Table 1 shows the baseline characteristic data of the general and spinal anesthesia groups. Before propensity score matching, the variables of age, sex, BMI, age-adjusted CCI score, fracture type, surgery type, surgical cost, surgery timing, frequency of sedation, frequency of delirium, hospital scale, year, ADLs before hospital admission, and some comorbidities were unbalanced between the two groups. After propensity score matching, the distributions of the patient background variables were well balanced between the general and spinal anesthesia groups.

Table 1.

Demographic and clinical characteristics of patients who underwent hip fracture surgery.

	Before propensity score matching			After propensity score matching
	General	Spinal	Standardized difference	General	Spinal	Standardized difference
n	11,673	8136		7519	7519
Age
65–79	3178 (27.2%)	2054 (25.2%)	0.045	1931 (25.7%)	1916 (25.5%)	0.005
80–89	5678 (48.6%)	3842 (47.2%)	0.028	3619 (48.1%)	3594 (47.8%)	0.006
90+	2817 (24.1%)	2240 (27.5%)	0.078	1969 (26.2%)	2009 (26.7%)	0.011
Sex
Male	2531 (21.7%)	1597 (19.6%)	0.052	1496 (19.9%)	1488 (19.8%)	0.003
Female	9142 (78.3%)	6539 (80.4%)	6023 (80.1%)	6031 (80.2%)
Body mass index
<18.5	2976 (25.5%)	2230 (27.4%)	0.043	2071 (27.5%)	2034 (27.1%)	0.009
18.5–24.9	6324 (54.2%)	4029 (49.5%)	0.094	3808 (50.6%)	3824 (50.9%)	0.006
25+	1083 (9.3%)	681 (8.4%)	0.032	642 (8.5%)	636 (8.5%)	0
Unknown	1290 (11.1%)	1196 (14.7%)	0.108	998 (13.3%)	1025 (13.6%)	0.009
Age-adjusted Charlson comorbidity index
2,3	1798 (15.4%)	1265 (15.5%)	0.003	1180 (15.7%)	1165 (15.5%)	0.006
4,5	7662 (65.6%)	5707 (70.1%)	0.096	5201 (69.2%)	5229 (69.5%)	0.007
6+	2213 (19.0%)	1164 (14.3%)	0.126	1138 (15.1%)	1125 (15.0%)	0.003
Smoking
Smoker	1012 (8.7%)	702 (8.6%)	0.004	652 (8.7%)	655 (8.7%)	0
Non-smoker	9328 (79.9%)	6442 (79.2%)	0.017	5946 (79.1%)	5964 (79.3%)	0.005
Unknown	1333 (11.4%)	992 (12.2%)	0.025	921 (12.2%)	900 (12.0%)	0.006
Delirium
-	11,539 (98.9%)	8067 (99.2%)	0.031	7449 (99.1%)	7453 (99.1%)	0
+	134 (1.1%)	69 (0.8%)		70 (0.9%)	66 (0.9%)
Fracture type
Transcervical	5551 (47.6%)	3655 (44.9%)	0.054	3357 (44.6%)	3355 (44.6%)	0
Intertrochanteric	5810 (49.8%)	4284 (52.7%)	0.058	3973 (52.8%)	3977 (52.9%)	0.002
Subtrochanteric	312 (2.7%)	197 (2.4%)	0.019	189 (2.5%)	187 (2.5%)	0
Surgery type
Arthroplasty	3999 (34.3%)	2263 (27.8%)	0.141	2205 (29.3%)	2182 (29.0%)	0.007
Other	7674 (65.7%)	5873 (72.2%)		5314 (70.7%)	5337 (71.0%)
Surgical cost (US$)
10,000 +	635 (5.4%)	156 (1.9%)	0.187	177 (2.4%)	156 (2.1%)	0.02
6,000–10,000	3525 (30.2%)	2168 (26.6%)	0.080	2101 (27.9%)	2084 (27.7%)	0.004
2000–6000	6994 (59.9%)	5624 (69.1%)	0.193	5059 (67.3%)	5092 (67.7%)	0.009
<2000	519 (4.4%)	188 (2.3%)	0.117	182 (2.4%)	187 (2.5%)	0.006
Timing of surgery
0 days	1822 (15.6%)	951 (11.7%)	0.114	900 (12.0%)	922 (12.3%)	0.009
1 day	3893 (33.4%)	2650 (32.6%)	0.017	2493 (33.2%)	2489 (33.1%)	0.002
2 days	3212 (27.5%)	2381 (29.3%)	0.040	2192 (29.2%)	2181 (29.0%)	0.004
3 days	2746 (23.5%)	2154 (26.5%)	0.069	1934 (25.7%)	1927 (25.6%)	0.002
Sedation
Pre and post	648 (5.6%)	766 (9.4%)	0.145	570 (7.6%)	585 (7.8%)	0.008
Pre or post	3287 (28.2%)	2519 (31.0%)	0.061	2303 (30.6%)	2321 (30.9%)	0.007
-	7738 (66.3%)	4851 (59.6%)	0.139	4646 (61.8%)	4613 (61.4%)	0.008
Hospital scale (bed#)
400+	8003 (68.6%)	4935 (60.7%)	0.166	4746 (63.1%)	4715 (62.7%)	0.008
<400	3670 (31.4%)	3201 (39.3%)		2773 (36.9%)	2804 (37.3%)
Year
2010.Apr–2013.Mar	2441 (20.9%)	2112 (26.0%)	0.121	1772 (23.6%)	1828 (24.3%)	0.016
2013.Apr–2016.Mar	3744 (32.1%)	2460 (30.2%)	0.041	2319 (30.8%)	2316 (30.8%)	0
2016.Apr–2019.Mar	5488 (47.0%)	3564 (43.8%)	0.064	3428 (45.6%)	3375 (44.9%)	0.014
Activity of daily living (eating) at hospital admission
Total depend	4327 (37.1%)	2429 (29.9%)	0.096	2387 (31.7%)	2357 (31.3%)	0.009
Partial depend	4580 (39.2%)	3380 (41.5%)	0.047	3113 (41.4%)	3088 (41.1%)	0.006
Independent	2542 (21.8%)	2110 (25.9%)	0.153	1836 (24.4%)	1894 (25.2%)	0.019
Unknown	224 (1.9%)	217 (2.7%)	0.053	183 (2.4%)	180 (2.4%)	0
Activity of daily living (grooming) at hospital admission
Dependent	9932 (85.1%)	6705 (82.4%)	0.073	6248 (83.1%)	6237 (82.9%)	0.005
Independent	1506 (12.9%)	1249 (15.4%)	0.072	1099 (14.6%)	1120 (14.9%)	0.008
Unknown	235 (2.0%)	182 (2.2%)	0.014	172 (2.3%)	162 (2.2%)	0.007
Activity of daily living (toileting) at hospital admission
Dependent	10,462 (89.6%)	7166 (88.1%)	0.048	6655 (88.5%)	6650 (88.4%)	0.003
Independent	926 (7.9%)	803 (9.9%)	0.070	697 (9.3%)	717 (9.5%)	0.007
Unknown	285 (2.4%)	167 (2.1%)	0.020	167 (2.2%)	152 (2.0%)	0.014
Activity of daily living (bathing) at hospital admission
Dependent	9632 (82.5%)	6702 (82.4%)	0.003	6199 (82.4%)	6200 (82.5%)	0.003
Independent	815 (7.0%)	714 (8.8%)	0.067	628 (8.4%)	632 (8.4%)	0
Unknown	1226 (10.5%)	720 (8.8%)	0.058	692 (9.2%)	687 (9.1%)	0.003
Activity of daily living (walking) at hospital admission
Dependent	9704 (83.1%)	6792 (83.5%)	0.011	6252 (83.1%)	6266 (83.3%)	0.005
Independent	800 (6.9%)	662 (8.1%)	0.046	592 (7.9%)	604 (8.0%)	0.004
Unknown	1169 (10.0%)	682 (8.4%)	0.055	675 (9.0%)	649 (8.6%)	0.014
Atrial fibrillation, atrial flutter (I48)
-	11,227 (96.2%)	7957 (97.8%)	0.094	7343 (97.7%)	7340 (97.6%)	0.007
+	446 (3.8%)	179 (2.2%)		176 (2.3%)	179 (2.4%)
Cerebrovascular diseases (I60–69)
-	10,625 (91.0%)	7731 (95.0%)	0.157	7130 (94.8%)	7115 (94.6%)	0.009
+	1048 (9.0%)	405 (5.0%)		389 (5.2%)	404 (5.4%)
Coagulopathy (D65–69)
-	11,565 (99.1%)	8103 (99.6%)	0.062	7480 (99.5%)	7486 (99.6%)	0.015
+	108 (0.9%)	33 (0.4%)		39 (0.5%)	33 (0.4%)
Mental disorders (F00–99)
-	8941 (76.6%)	6156 (75.7%)	0.021	5714 (76.0%)	5713 (76.0%)	0
+	2732 (23.4%)	1980 (24.3%)		1805 (24.0%)	1806 (24.0%)
Neurologic diseases (G00–99)
-	9089 (77.9%)	6692 (82.3%)	0.110	6132 (81.6%)	6153 (81.8%)	0.005
+	2584 (22.1%)	1444 (17.7%)		1387 (18.4%)	1366 (18.2%)
Respiratory diseases (J00–99)
-	10,767 (92.2%)	7562 (92.9%)	0.027	6968 (92.7%)	6990 (93.0%)	0.012
+	906 (7.8%)	574 (7.1%)		551 (7.3%)	529 (7.0%)
Thrombophlebitis (I80)
-	11,470 (98.3%)	8023 (98.6%)	0.024	7413 (98.6%)	7412 (98.6%)	0
+	203 (1.7%)	113 (1.4%)		106 (1.4%)	107 (1.4%)
Vertebral diseases (M40–54)
-	11,311 (96.9%)	7966 (97.9%)	0.063	7339 (97.6%)	7351 (97.8%)	0.013
+	362 (3.1%)	170 (2.1%)		180 (2.4%)	168 (2.2%)

Total Depend: Total Dependent; Partial Depend: Partial dependent; Surgical Cost was calculated at a rate of 1US$ = 100 Yen.

After propensity score matching, the frequencies of aspiration pneumonia in the general and spinal anesthesia groups were 1.8% and 1.5%, respectively (p = 0.158). The frequencies of other pulmonary complications (pneumonia, bronchitis, atelectasis, pneumothorax, pulmonary edema, and pulmonary embolism) in the general and spinal anesthesia groups were 1.5% and 1.5%, respectively (p = 0.893). The mortalities in the general and spinal anesthesia groups were 1.4% and 1.2%, respectively (p = 0.219). The type of anesthesia was not associated with the occurrences of aspiration pneumonia, other pulmonary complications, or mortality in the hospital (Table 2). Of the 15,038 patients who matched the propensity score, 246 had aspiration pneumonia. The percentage of patients who always required assistance with eating before admission was 48% in patients with aspiration pneumonia and 31% in patients without aspiration pneumonia (p < 0.01).

Table 2.

Associations between anesthetic methods and outcomes.

Outcomes (ICD-10 code)	General (n =7519)	Spinal (n =7519)	Logistic regression analysis		Instrumental variable analysis
			OR (95%CI)	p value	OR (95%CI)	p Value
Aspiration pneumonia (J69.0)	1.8% (134)	1.5% (112)	1.200 (0.932–1.545)	0.158	1.289 (0.964–1.724)	0.086
Other pulmonary complications	1.5% (114)	1.5% (112)	1.018 (0.783–1.324)	0.893	0.985 (0.731–1.328)	0.922
Pneumonia (J13–18)	1.0% (76)	0.9% (67)	1.136 (0.816–1.580)	0.450	1.156 (0.78–1.712)	0.470
Bronchitis (J20–22)	0.2% (14)	0.2% (13)	1.077 (0.506–2.293)	0.847	0.91 (0.44–1.882)	0.799
Atelectasis (J981)	0% (2)	0% (3)	0.667 (0.111–3.990)	0.657	1.066 (0.19–5.984)	0.942
Pnemothorax (J93)	0% (1)	0.1% (4)	0.250 (0.028–2.236)	0.215	0.322 (0.069–1.594)	0.168
Pulmonary edema (J81)	0.1% (5)	0% (3)	1.667 (0.398–6.978)	0.484	2.763 (0.65–11.737)	0.169
Pulmonary embolism (I26)	0.3% (19)	0.3% (22)	0.863 (0.467–1.596)	0.639	0.695 (0.353–1.367)	0.292
Death	1.4% (105)	1.2% (88)	1.196 (0.899–1.590)	0.219	1.053 (0.733–1.513)	0.780

Instrumental variable analyses using the ratio of general anesthesia and spinal anesthesia performed at each hospital between 1 April 2019 and 31 March 2020 as an instrumental variable were performed, and odds ratio (OR) and their 95% confidence interval (CI) were determined using a two-stage, residual inclusion model.

According to the instrumental variable analysis results, the mean proportion of patients who received general anesthesia was 80.0% among those admitted to hospitals with higher proportions of general anesthesia and 28.6% among those admitted to hospitals with lower proportions of general anesthesia. The F-statistic was 4,427, suggesting a strong instrument. In the two-stage residual inclusion model, the anesthesia method was not associated with aspiration pneumonia (p = 0.086). The anesthesia method was also not associated with other pulmonary complications (p = 0.922) or death (p = 0.780) (Table 2).

The results of the multivariable logistic regression analysis showed that advanced age, male sex, lower BMI, cerebrovascular disease, dementia, and ADL dependency (eating) were associated with the occurrence of aspiration pneumonia (Table 3). Compared with patients in their 60s or 70s, patients in their 80s and ≥90s had nearly 1.6-fold and 2.3-fold higher odds of aspiration pneumonia, respectively. Male patients had more than 3-fold higher odds of aspiration pneumonia compared with female patients. Patients with a BMI <18.5 kg/m² had 1.5-fold higher odds of aspiration pneumonia compared with patients with a normal BMI. Compared with patients without cerebrovascular disease and dementia, patients with cerebrovascular disease and dementia had nearly 1.7-fold and 1.9-fold higher odds of aspiration pneumonia, respectively. However, Parkinson disease/syndrome and anesthesia types were not associated with the occurrence of aspiration pneumonia. Patients always needing eating assistance had about 1.8-fold higher odds of aspiration pneumonia compared with patients with the ability to eat independently.

Table 3.

Multivariable analysis: Factors affecting aspiration pneumonia after elderly hip fracture surgery.

Variables		B	OR	95% CI	p value
Age	90+	0.834	2.303	1.635–3.246	<0.001
	80–89	0.451	1.570	1.145–2.154	0.005
	60–79	-	1
Sex	Male	1.309	3.703	2.937–4.669	<0.001
	Female		1
Body mass index (kg/m²)	25+	−0.475	0.622	0.365–1.061	0.081
	<18.5	0.430	1.538	1.219–1.941	<0.001
	18.5–24.9		1
Cerebrovascular Disease (I60–69)	+	0.504	1.655	1.212–2.260	0.002
	-		1
Parkinson disease and syndrome (G20–22)	+	−0.065	0.937	0.436–2.014	0.869
	-		1
Dementia (F00,01,02,03,F051,G30,G311)	+	0.639	1.895	1.482–2.424	<0.001
	-		1
Activity of daily living (eating)	Total dependent	0.570	1.768	1.252–2.495	0.001
	Partial dependent	0.339	1.404	0.991–1.987	0.056
	Independent		1
Anesthesia	General	0.181	1.198	0.944–1.521	0.136
	Spinal		1

B: Logistic regression coefficient; OR: odds ratio; CI: confidence interval. The alphabets and numbers in parentheses mean the ICD-10 code.

Discussion

To our knowledge, three groups have investigated aspiration pneumonia in hip fracture surgery.^20,21,24 In two of the three studies,^20,24 the incidence of aspiration pneumonia was not related to the anesthesia method, which is consistent with our results. In the third study,²¹ the effects of the anesthesia method in the development of aspiration pneumonia were not investigated. The frequency of aspiration pneumonia is affected by the diagnosis method (e.g., a radiologic study). One large study²⁴ used the diagnosis code like we did, and the frequency of aspiration pneumonia occurrence in that study was similar to the frequency of aspiration pneumonia occurrence in our study.

Many studies have not focused on aspiration pneumonia and instead have investigated the relationships between different anesthesia methods and pulmonary complications and mortality in hip surgery.^24–28 By looking at the research target years of these studies in chronological order, it can be seen that the evaluation of anesthesia methods has changed since around 2010. Older retrospective studies have shown that general anesthesia has higher incidences of pulmonary complications and mortality than local anesthesia.^24–26 However, recent retrospective studies have shown that neither method of anesthesia has a significant effect on mortality or pulmonary complications.^27,28 Furthermore, recent prospective studies have shown no significant association between the anesthesia method and the incidence of pneumonia or mortality in hip fracture surgery.^29,30 The results of the present study are consistent with the results of these recent reports. In the present study, the frequencies of atelectasis, pneumothorax, and pulmonary edema were very low, and there were no significant differences between the anesthesia groups. We could not find any previous reports mentioning these complications, probably because of their low frequencies. Regarding pulmonary embolism, two studies reported that there was no significant difference due to the anesthesia method.^22,31

There are several possible reasons why there was no significant difference in the incidence of aspiration pneumonia between the general and spinal anesthesia groups in this study. First, the present study showed that the eating dependence rates for patients with and without aspiration pneumonia were 48% and 31%, respectively: that is, patients who had eaten dependently before admission were more likely to develop aspiration pneumonia. Preoperative ADL abilities (especially eating) are important for maintaining the airway protective reflex and swallowing function.³² Adjusting for preoperative ADL independence may be one of the reasons why there was no significant difference in the incidence of aspiration pneumonia between the two groups in this study. Second, it was presumed that the start time of oral intake of water and food after surgery was adjusted according to each patient’s condition. Such clinical adjustments may be another reason why there was no significant difference in the incidence of aspiration pneumonia between the two groups. Third, the anesthetics used in recent years are eliminated from the body extremely quickly.^11–13 Sevoflurane and desflurane allow for rapid recovery from anesthesia due to their low blood and tissue solubility.^11,12,33 Even elderly people have fewer than 15 min to extubate and fewer than about 20 min to recover orientation (location, time, and date).³⁴ Remifentanil is rapidly hydrolyzed by blood and tissue esterases and becomes metabolites with very low potency, so the action of remifentanil disappears in about 10 min, even after long-term use, regardless of liver function and age.^13,35 In Japan, sevoflurane has been clinically available since the 1990s, while remifentanil has been available since 2007, and desflurane since 2011. Residual neuromuscular blockers are associated with an increased risk of postoperative pulmonary complications.^36,37 Sugammadex, which can antagonize muscle relaxants quickly and completely compared to traditional anticholinesterases, has been in clinical use in Japan since 2010 and is currently used in more than 90% of cases.^38–40 Since this study used data from 2010 to 2019, it might be inferred that these short-acting anesthetics and sugammadex reduced the incidence of aspiration pneumonia after general anesthesia. The use of short-acting anesthetics and sugammadex is widespread worldwide. The results of recent retrospective and prospective studies may support this third reason.^27–30 Fourth, many elderly patients with hip fractures are frail and suffer from cognitive impairment or chronic constipation. In frail patients, the muscles used to exhale vomit are presumed to be weak. Past literature has asserted that coughing and swallowing are controlled primarily by the brainstem; however, recent studies have reported that the cortical and subcortical structures play critical roles in the airway protective reflex.^41,42 If this theory is correct, then patients with cognitive impairments, whose brains are atrophying, are likely to be prone to aspiration. Furthermore, chronic constipation is associated with ileus.⁴³ If ileus occurs, there is an increased risk of vomiting. Therefore, hip fracture patients are speculated to be more likely to develop aspiration pneumonia due to their deteriorating physical strength and their health conditions.^43-45 These overwhelming factors may have negated the slight difference between anesthesia methods and might have resulted in the lack of a significant difference in the incidence of aspiration pneumonia between the general and spinal anesthesia groups in this study.

In the multivariable analysis, the age and BMI results were consistent with a previous report that older age and lower BMI are associated with the development of aspiration pneumonia.^1,2,20–22 So far, no consistent results have been reported regarding the effects of sex.^1,2,21 The distance between the hyoid bone and the mandible increases with aging more in males than in females.⁴⁶ This change in the position of the hyoid bone indicates atrophy of the surrounding muscles, and it narrows the space in the larynx, impairing safe swallowing movements. We think that this may be one of the reasons why males had more occurrences of aspiration pneumonia than females. When it comes to the presence of Parkinson disease/syndrome, it is unclear why there was no significant difference in the occurrence of aspiration pneumonia. The occurrence of a hip fracture means that the patient was able to walk or stand. At that stage, patients with Parkinson disease/syndrome may still have relatively mild dysphagia. Assessment of swallowing ability and intervention by a speech therapist can reduce postoperative pneumonia.^47,48 These techniques should be used to prevent aspiration pneumonia, especially in high-risk patients: male sex, ≥ 80 years of age, low BMI, cerebrovascular disease, dementia, and need for eating assistance.

Several limitations should be noted. First, the results of the present study are limited by the retrospective study design. Since only in-hospital information was available, complications after discharge were not involved. However, Japanese insurance systems allow patients to stay in the hospital longer than North American systems. In the present study, the mean postoperative hospitalization period was 27 ± 19 days, which was longer than the period of 7 ± 10 days in the study by Studer et al.³; therefore, we do not think the observation period was short. Second, we did not know when aspiration pneumonia developed because the days of disease onset were not recorded in the DPC database. Therefore, to exclude cases in which aspiration pneumonia developed before surgery, we included only cases in which surgery was performed within 4 days, including the day of hospitalization; we assumed that if patients were expected to develop aspiration pneumonia, then their hip fracture surgeries were postponed and treatment for aspiration pneumonia was prioritized. Furthermore, the instrumental variable method is a method of using instrumental variables to perform a retrospective observational study similar to a randomized controlled trial.⁴⁹ The results of both the propensity score analysis and the instrumental variable method were similar, demonstrating the robustness of the results of this study. However, prospective investigations are needed because this study could not determine whether aspiration pneumonia occurred soon after surgery or a few days later. Third, some patients with femoral fractures cannot undergo femoral surgery immediately after hospitalization because they are undergoing antithrombotic therapy or have problems such as cardiac function problems, poor control of diabetes mellitus, or concomitant trauma. Excluding these patients in this study may have resulted in low rates of aspiration pneumonia and mortality. Fourth, as the study population was primarily Japanese, the results may not be generalizable to people in other countries.

Conclusion

The incidence of aspiration pneumonia was not significantly different between general and spinal anesthesia in elderly hip fracture surgery patients. The incidence of aspiration pneumonia was 1.8% (total of general and spinal anesthesia). The study excluded critically ill patients who could not undergo surgery immediately after admission; the actual incidence of aspiration pneumonia may have exceeded 1.8%. Predictive factors of aspiration pneumonia were advanced age, male sex, lean body, cerebrovascular disease, dementia, and ADL dependency (eating). Regardless of the anesthesia method, great care should be taken to prevent postoperative aspiration pneumonia, especially in elderly patients with these predictive factors.

Footnotes

Acknowledgements

The authors thank Chiharu Okada, MD, PhD, MBA (Department of Hospital Management Assistance, National Organization Headquarters, Tokyo, Japan) for arranging this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Taeko Fukuda

References

Kozlow

Berenholtz

Garrett

, et al. Epidemiology and impact of aspiration pneumonia in patients undergoing surgery in Maryland, 1999-2000. Crit Care Med 2003; 31: 1930–1937.

Fineberg

Oglesby

Patel

, et al. Incidence, risk factors, and mortality associated with aspiration in cervical spine surgery. Spine 2013; 38: E1189–E1195.

Studer

Räber

Ott

, et al. Risk factors for fatal outcome in surgical patients with postoperative aspiration pneumonia. Int J Surg 2016; 27: 21–25.

Teramoto

Fukuchi

Sasaki

, et al. High incidence of aspiration pneumonia in community- and hospital-acquired pneumonia in hospitalized patients: a multicenter, prospective study in Japan. J Am Geriatr Soc 2008; 56: 577–579.

Johnell

Kanis

. An estimate of the worldwide prevalence and disability associated with osteoporotic fractures. Osteoporos Int 2006; 17: 1726–1733.

Gullberg

Johnell

Kanis

. World-wide projections for hip fracture. Osteoporos Int 1997; 7: 407–413.

Sinclair

Téllez

Tapia

, et al. Contralateral R1 and R2 components of the laryngeal adductor reflex in humans under general anesthesia. The Laryngoscope 2017; 127: E443–E448.

Tanaka

Isono

Ishikawa

, et al. Laryngeal reflex before and after placement of airway interventions: endotracheal tube and laryngeal mask airway. Anesthesiology 2005; 102: 20–25.

Erskine

Murphy

Langton

, et al. Effect of age on the sensitivity of upper airway reflexes. Br J Anaesth 1993; 70: 574–575.

10.

Dua

Surapaneni

Kuribayashi

, et al. Effect of aging on hypopharyngeal safe volume and the aerodigestive reflexes protecting the airways. The Laryngoscope 2014; 124: 1862–1868.

11.

Behne

Wilke

H-J

Harder

. Clinical pharmacokinetics of sevoflurane. Clin Pharmacokinet 1999; 36: 13–26.

12.

Yasuda

Lockhart

Eger

II , et al. Kinetics of desflurane, isoflurane, and halothane in humans. Anesthesiology 1991; 74: 489–498.

13.

Westmoreland

Hoke

Sebel

, et al. Pharmacokinetics of remifentanil (GI87084B) and its major metabolite (GI90291) in patients undergoing elective inpatient surgery. Anesthesiology 1993; 79: 893–903.

14.

World Health Organization . Global Database on Body Mass Index, . Available from: http://www.apps.who.int/bmi/index.jsp?introPage=intro_3.html.BMI Classification

15.

Collin

Wade

Davies

, et al. The barthel ADL index: a reliability study. Int Disabil Stud 1988; 10: 61–63.

16.

Fukuda

Imai

Nakadera

, et al. Postoperative daily living activities of geriatric patients administered general or spinal anesthesia for hip fracture surgery: a retrospective cohort study. J Orthopaedic Surg 2018; 26: 230949901775410. DOI: 10.1177/2309499017754106.

17.

Quan

Sundararajan

Halfon

, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005; 43: 1130–1139.

18.

Charlson

Szatrowski

Peterson

, et al. Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47: 1245–1251.

19.

Sjolander

Martinussen

. Instrumental variable estimation with the R package ivtools. Epidemiologic Methods 2019; 8: 20180024. DOI: 10.1515/em-2018-0024.

20.

Byun

Shon

Kim

, et al. Risk factors and prognostic implications of aspiration pneumonia in older hip fracture patients: a multicenter retrospective analysis. Geriatr Gerontology International 2019; 19: 119–123.

21.

Higashikawa

Shigemoto

Goshima

, et al. Risk factors for the development of aspiration pneumonia in elderly patients with femoral neck and trochanteric fractures. Medicine 2020; 99: e19108. DOI: 10.1097/MD.0000000000019108.

22.

Fields

Dieterich

Buterbaugh

, et al. Short-term complications in hip fracture surgery using spinal versus general anaesthesia. Injury 2015; 46: 719–723.

23.

Sjolander

Dahlqwist

Martinussen

. Ivtools: instrumental variables. R package version 2.3.0, https://CRAN.R-project.org/package=ivtools (2020).

24.

Newman

Silber

Elkassabany

, et al. Comparative effectiveness of regional versus general anesthesia for hip fracture surgery in adults. Anesthesiology 2012; 117: 72–92.

25.

Radcliff

Henderson

Stoner

, et al. Patient risk factors, operative care, and outcomes among older community-dwelling male veterans with hip fracture. The J Bone Jt Surgery-American 2008; 90: 34–42.

26.

Chu

C-C

Weng

S-F

Chen

K-T

, et al. Propensity score-matched comparison of postoperative adverse outcomes between geriatric patients given a general or a neuraxial anesthetic for hip surgery. Anesthesiology 2015; 123: 136–147.

27.

Basques

Bohl

Golinvaux

, et al. General versus spinal anaesthesia for patients aged 70 years and older with a fracture of the hip. Bone Jt J 2015; 97-B: 689–695.

28.

Gremillet

Jakobsson

. Acute hip fracture surgery anaesthetic technique and 30-day mortality in sweden 2016 and 2017: a retrospective register study. F1000Research 2018; 7: 1009. DOI: 10.12688/f1000research.15363.2.

29.

Zheng

Tan

Gao

, et al. Comparative efficacy of neuraxial and general anesthesia for hip fracture surgery: a meta-analysis of randomized clinical trials. BMC Anesthesiology 2020; 20: 162. DOI: 10.1186/s12871-020-01074-y.

30.

Shin

Kim

Park

, et al. Effects of anesthesia techniques on outcomes after hip fracture surgery in elderly patients: a prospective, randomized, controlled trial. J Clin Med 2020; 9: 1605. DOI: 10.3390/jcm9061605.

31.

Whiting

Molina

Greenberg

, et al. Regional anaesthesia for hip fracture surgery is associated with significantly more peri-operative complications compared with general anaesthesia. Int Orthopaedics 2015; 39: 1321–1327.

32.

Nakayama

Tohara

Hino

, et al. The effects of ADL on recovery of swallowing function in stroke patients after acute phase. J Oral Rehabil 2014; 41: 904–911.

33.

Sakai

Connolly

Klauck

. Inhalation anesthesiology and volatile liquid anesthetics: focus on isoflurane, desflurane, and sevoflurane. Pharmacotherapy 2005; 25: 1773–1788.

34.

Heavner

Kaye

Lin

B-K

, et al. Recovery of elderly patients from two or more hours of desflurane or sevoflurane anaesthesia †. Br J Anaesth 2003; 91: 502–506.

35.

Scott

Perry

. Remifentanil: a review of its use during the induction and maintenance of general anesthesia. Drugs 2005; 65: 1793–1823.

36.

Stankiewicz-Rudnicki

. Neuromuscular blockade in the elderly. Anestezjologia Intensywna Terapia 2016; 48: 257–260.

37.

Eikermann

Groeben

Bünten

, et al. Fade of pulmonary function during residual neuromuscular blockade. Chest 2005; 127: 1703–1709.

38.

Ledowski

Szabó-Maák

Loh

, et al. Reversal of residual neuromuscular block with neostigmine or sugammadex and postoperative pulmonary complications: a prospective, randomised, double-blind trial in high-risk older patients. Br J Anaesth 2021; 127: 316–323.

39.

Kheterpal

Vaughn

Dubovoy

, et al. Sugammadex versus neostigmine for reversal of neuromuscular blockade and postoperative pulmonary complications (STRONGER). Anesthesiology 2020; 132: 1371–1381.

40.

Orihara

Takazawa

Horiuchi

, et al. Comparison of incidence of anaphylaxis between sugammadex and neostigmine: a retrospective multicentre observational study. Br J Anaesth 2020; 124: 154–163.

41.

Ebihara

Kohzuki

. Effect of aging on cough and swallowing reflexes: implications for preventing aspiration pneumonia. Lung 2012; 190: 29–33.

42.

Michou

Hamdy

. Cortical input in control of swallowing. Curr Opin Otolaryngol Head Neck Surg 2009; 17: 166–171.

43.

Lee

Hong

, et al. Risk factors for postoperative ileus following orthopedic surgery: the role of chronic constipation. J Neurogastroenterology Motil 2015; 21: 121–125.

44.

Patel

Brennan

, et al. Association of a modified frailty index with mortality after femoral neck fracture in patients aged 60 years and older. Clin Orthopaedics Relat Res 2014; 472: 1010–1017.

45.

Seitz

Adunuri

Gill

, et al. Prevalence of dementia and cognitive impairment among older adults with hip fractures. J Am Med Directors Assoc 2011; 12: 556–564.

46.

Feng

Todd

, et al. Age-related changes of hyoid bone position in healthy older adults with aspiration. The Laryngoscope 2014; 124: E231–E236.

47.

Chojin

Kato

Noguchi

, et al. Evaluating a novel swallowing assessment as a predictor of mortality and recurring pneumonia in elderly patients with pneumonia. Respir Invest 2021; 59: 783–791.

48.

Starks

Harbert

. Aspiration prevention protocol: decreasing postoperative pneumonia in heart surgery patients. Crit Care Nurse 2011; 31: 38–45.

49.

Aso

Yasunaga

. Introduction to instrumental variable analysis. Ann Clin Epidemiol 2020; 2: 69–74. doi: 10.37737/ace.2.3_69

Aspiration pneumonia and anesthesia techniques in hip fracture surgery in elderly patients: A retrospective cohort study using administrative data