Long-Term Survival of Patients Receiving Artificial Nutrition in Japanese Psychiatric Hospitals.

BACKGROUND/AIMS: Most patients with dementia suffer from dysphagia in the terminal stage of the disease. In Japan, most elderly patients with dysphagia receive either tube feeding or total parenteral nutrition.
METHODS: In this study, we investigated the factors determining longer survival with artificial nutrition. Various clinical characteristics of 168 inpatients receiving artificial nutrition without oral intake in psychiatric hospitals in Okayama Prefecture, Japan, were evaluated.
RESULTS: Multiple logistic regression analysis showed that the duration of artificial nutrition was associated with a percutaneous endoscopic gastrostomy (PEG) tube, diagnosis of mental disorder, low MMSE score, and absence of decubitus.
CONCLUSION: Patients with mental disorders survived longer than those with dementia diseases on artificial nutrition. A PEG tube and good nutrition seem to be important for long-term survival.

Introduction

The aged population has increased in many developed countries including Japan. The frequency of dementia rises with advancing age. Therefore, the number of patients with dementia grows dramatically in many countries. Most patients with dementia suffer from dysphagia in the terminal stage of the disease. A recent systematic review reported that there was no evidence of longer survival of dementia patients receiving enteral tube feeding [1]. However, in many geriatric facilities worldwide, including Japan, clinical physicians continue to provide tube feeding to dementia patients for traditional, ethical, and administrative reasons [2,3]. Especially in Japan, most elderly patients with dysphagia (about 90%) receive either tube feeding or total parenteral nutrition (TPN) [4].

As stated above, several studies of the survival rate of dementia patients receiving percutaneous endoscopic gastrostomy (PEG) have shown that artificial nutrition by PEG tube feeding did not improve the prognosis of dementia patients compared to hand feeding [1,5,6,7]. However, in those studies, there was no homogeneity in the dementia patients with and without PEG tube feeding. It is most likely that many patients without feeding tubes in those studies had better health [8] and that a so-called selection bias affected the results. Recently, some studies have shown that dementia patients with PEG tubes survived longer than those without [2,4,8]. Furthermore, it was also reported that dementia patients receiving PEG tube feeding survived as long as patients without dementia receiving PEG tube feeding [9,10,11,12]. In a recent report from Japan, 66% of 931 consecutive dementia patients with PEG tube feeding survived over 1 year, and more than 50% of those patients survived over 2 years [3]. In another report, 74 patients (39.6% of 187 patients) had been receiving artificial nutrition for more than 2 years and 28 patients (15.0%) for more than 5 years [13].

Recently, several studies have reported factors affecting the duration of survival after PEG tube insertion. Younger age, female sex, higher albumin level, higher lymphocyte counts, and absence of cardiac failure were reported to predict longer survival among patients with a PEG tube [3,9,10,11]. These studies did not include patients with psychiatric diseases. Currently, in Japan, not a few patients with psychiatric diseases suffer from dysphagia in the chronic stage and receive parenteral nutrition [13]. However, there have been no studies focusing on the factors affecting survival time after commencement of artificial nutrition, including in patients with mental diseases and patients receiving various methods of artificial nutrition. In this study, in order to determine the factors related to longer survival with artificial nutrition, we investigated the actual conditions of patients undergoing artificial nutrition by methods including nasogastric (NG) tube feeding, PEG tube feeding, and TPN in psychiatric hospitals in Okayama Prefecture, Japan.

Subjects and Methods

Artificial Nutrition

Artificial hydration and nutrition include enteral and intravenous nutrition. Enteral nutrition mainly consists of NG tube and PEG feeding, while intravenous nutrition comprises peripheral venous nutrition and TPN. Needless to say, peripheral venous nutrition cannot sustain human life for long. Therefore, only patients with a PEG tube, NG tube, or TPN and without oral intake were included in this study.

Subjects

This was a cross-sectional observational study. Almost all wards for dementia with behavioral and psychological symptoms of dementia are situated in psychiatric hospitals in Japan. Due to the explosive increase in aged dementia patients, the number of care facilities for patients with dementia is not sufficient. Therefore, quite a few patients with dementia are cared for in many psychiatric hospitals even after behavioral and psychological symptoms of dementia have been well controlled.

We asked the Association of Okayama Psychiatry Hospitals to participate in the survey of inpatients undergoing artificial feeding. There are 20 psychiatric hospitals in Okayama Prefecture, and 18 of them agreed to participate in this survey. Inpatients receiving artificial nutrition for more than 1 month as of February 17, 2014, were included. All inpatients receiving artificial nutrition were evaluated in 16 of 18 hospitals. In 2 hospitals, 33 of 58 patients and 6 of 15 patients were investigated. Reasons for nonparticipation were heavy workload and/or shortage of personnel in the wards.

On February 17, 2014, there were 4,101 inpatients (2,094 women and 2,007 men) in 18 psychiatric hospitals of Okayama Prefecture. Of them, 221 patients (5.4% of all inpatients) had been receiving artificial nutrition for more than 1 month. Of the 221 patients, 75 patients were male (3.7% of all male inpatients) and 146 were female (7.0% of all female inpatients). Of 221 patients receiving continuous artificial nutrition, 187 patients (130 women and 57 men, 84.6% of 221 patients) were fully investigated. Of 187 patients, 168 patients (114 women and 54 men, 89.8% of 187 patients) received artificial nutrition without oral intake.

Clinical Diagnosis

All patients with Alzheimer's disease (AD) met the criteria for probable AD formulated by the NIA-AA [14]. AD patients with severe vascular lesions were diagnosed as possible AD with vascular lesions according to the criteria of the NIA-AA [14]. All patients with vascular dementia (VaD) were diagnosed according to the criteria for probable VaD of the AHA-ASA [15]. All patients with dementia with Lewy bodies (DLB) were diagnosed according to the criteria for probable DLB of the DLB Consortium [16]. All patients with frontotemporal dementia met the international consensus criteria for probable behavioral variant frontotemporal dementia [17]. Patients with schizophrenia fulfilled the International Classification of Diseases (ICD)-10 criteria for F2 codes, and patients with mood disorders fulfilled the ICD-10 criteria for F3 codes. Other disorders were diagnosed according to ICD-10 criteria.

Questionnaires and Clinical Scales

Clinical characteristics including age, clinical diagnosis, and duration and methods of artificial nutrition were investigated. The presence or absence of physical restraints and pressure ulcers at the time of the survey was included, and the number of cases of pneumonia in the previous year was also recorded.

The Functional Independence Measure (FIM) is a widely used functional assessment tool [18]. The FIM contains 18 items (13 motor tasks and 5 cognitive tasks). Each item is objectively rated on a 7-point scale, and a lower score indicates more severe impairment. The Mini-Mental State Examination (MMSE) is a screening test that has been used widely to evaluate the cognitive state [19]. The total score of the MMSE ranges from 0 to 30. A lower score indicates more severe impairment.

Respondents

Questionnaires were completed for all subjects by trained nurses or geriatric psychiatrists who were well acquainted with and chiefly in charge of the participants being rated. All raters had daily contact with the individuals being studied. The MMSE was evaluated by geriatric psychiatrists, clinical psychologists, or trained occupational therapists.

Statistics

Statistical analyses were performed using PASW Statistics 18.0 (SPSS, Chicago, Ill., USA). The comparison between the three groups was made using one-way analysis of variance (ANOVA), followed by Tukey's post hoc test to determine significant differences between the means of the data groups. Comparison of proportions between the three groups was calculated using a χ2 test (2 × 3 table).

To identify which variables were significantly correlated with duration of artificial nutrition, we performed a multiple linear regression using duration of artificial nutrition as a dependent variable. Independent variables were gender, age, diagnosis (dementia = 1, other = 0), presence or absence of restraint and decubitus, number of incidents of pneumonia, FIM motor scores, FIM cognitive scores, MMSE scores, and methods of artificial nutrition. In a multiple regression analysis, three dummy variables were used instead of ‘methods of artificial nutrition’ (nominal scale). One dummy variable consisted of PEG = 1 with other = 0; the other two consisted of NG tube = 1 with other = 0 and TPN = 1 with other = 0. Subsequently, forward stepwise linear regression models (entrance criterion p = 0.05, exit p = 0.10) determined if disparate predictors accounted for unique additive variance in the duration of artificial nutrition. A value of p < 0.05 was accepted as significant.

Statement of Ethics

This study adhered to the 1975 Helsinki Declaration of Human Rights. The study protocol with a list of the attending psychiatric hospitals was approved by the Internal Ethical Committee of Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences.

Results

Demographic Characteristics

A total of 168 patients (114 women and 54 men, 89.8% of 187 patients) received artificial nutrition without oral intake. The mean age of the patients was 78.5 ± 11.1 years, and the mean duration of artificial nutrition was 30.6 ± 31.9 months (table 1). Sixty-six patients (39.3% of 168 patients) had been receiving artificial nutrition for more than 2 years and 27 patients (16.1%) for more than 5 years.

Table 1

Age, diagnosis, duration of artificial nutrition, and pneumonia (n = 168)

Characteristics	n	%
Age
≤49 years	2	1.2
50–59 years	6	3.6
60–69 years	28	16.7
70–79 years	42	25.0
80–89 years	64	38.0
≥90 years	26	15.5
Diagnosis
Probable AD	71	42.2
Schizophrenia	33	19.6
VaD	23	13.7
DLB	9	5.4
Suffocation accidents	9	5.4
Others	23	13.7
Duration
≤12 months	58	34.5
13–24 months	44	26.2
25–36 months	19	11.3
37–48 months	12	7.1
49 - 60 months	8	4.8
≥61 months	27	16.1
Pneumonia incidence
0	93	55.3
1	36	21.4
2	21	12.5
3	6	3.6
4	6	3.6
5 - 6	6	3.6

Clinical diagnoses are shown in table 1. Patients with probable AD (n = 71) formed the largest group, schizophrenia patients (n = 33) the second, and those with VaD (n = 23) the third. Among 71 AD patients, 26 (36.6%) had been receiving artificial nutrition for more than 2 years and 7 patients (9.9%) for more than 5 years. As for the methods of artificial nutrition, 90 patients were receiving it by PEG tube (53.6%), 63 patients by NG tube (37.5%), and 15 patients by TPN (8.9%). Most patients were not under any physical restraint (n = 148, 88.1%). Pressure ulcers were found in 22 patients (13.1%), and pneumonia in the previous year was recorded for 75 patients (44.6%).

The distribution of FIM scores is shown in table 2. Twenty-eight patients had the lowest score (13 points) on the FIM motor domain, and more than two thirds of patients (n = 124) scored 14. About 70% of the patients (n = 120) had the lowest score (5 points) on the FIM cognitive domain. On the FIM total score, 25 patients scored 18 (lowest score) and 93 patients (55.4%) scored 19. The mean MMSE score of 168 patients was 2.0 ± 5.0 points, and more than two thirds (n = 129) scored 0.

Table 2

Patients’ FIM and MMSE scores (n = 168)

Characteristics	n	%
Motor scores of FIM (13–91)
13	28	16.7
14–19	137	81.5
20–26	1	0.6
27–39	0	0
≥40	2	1.2
Cognitive scores of FIM (5–35)
5	120	71.4
6–7	17	10.1
8–10	17	10.1
11–15	7	4.2
≥16	7	4.2
Total scores of FIM (18–126)
18	25	14.9
19–27	134	79.7
28–36	5	3.0
37–54	2	1.2
≥55	2	1.2
MMSE score (0–30)
0	129	76.8
1–5	16	9.5
6–10	11	6.5
11–15	7	3.6
16–30	5	3.6

Comparison between Short, Medium, and Long Duration of Artificial Nutrition and Factors Related to Long Duration of Artificial Nutrition

Comparisons between patients with short, medium, and long duration of artificial nutrition are shown in table 3. The proportion of the three methods of artificial nutrition was significantly different between the three groups. Pneumonia in the previous year was more frequent in patients with a short duration of artificial nutrition. The mean MMSE scores and the percentage of physical restraints in the three groups showed close-to-significant differences (table 3).

Table 3

Comparison of patients with a different duration of artificial nutrition

	Duration of artificial nutrition			F	p
	short (≤12 months)	middle (13–36 months)	long (≥37 months)
Patients, n	58	63	47
Mean age ± SD, years	78.8 ± 10.9	78.1 ± 10.3	78.7 ± 12.5	0.085	0.919
Mean duration ± SD, months	6.0 ± 3.3	21.7 ± 6.3	72.9 ± 30.6	221.926	<0.001
Mean MMSE ± SD	2.9 ± 6.1	1.9 ± 4.4	0.8 ± 2.7	2.598	0.077
Mean FIM motor scores ± SD	14.4 ± 3.7	15.0 ± 6.1	13.8 ± 0.4	0.936	0.394
Mean FIM cognitive scores ± SD	7.1 ± 5.2	6.4 ± 3.3	5.8 ± 1.9	1.408	0.248
Mean FIM total scores ± SD	21.5±7.9	21.4 ± 8.2	19.6 ± 2.0	1.137	0.323
				χ2	p
Men/women	21/37	20/43	13/34	0.877	0.645
Dementia/others	43/15	43/20	27/20	3.330	0.189
PEG/NG tube/TPN	29/20/9	23/35/5	38/8/1	26.289	<0.001
Pneumonia (+/–)	37/21	25/38	13/34	14.720	0.001
Restraint (+/–)	9/49	10/53	1/46	5.952	0.051
Decubitus (+/–)	12/46	6/57	4/43	4.514	0.105

SD = Standard deviation.

Multiple logistic regression analysis showed that the duration of artificial nutrition was predicted by the use of PEG tube, diagnosis, MMSE score, and decubitus according to the following formula (R2 = 0.172) (table 4): duration of artificial nutrition (months) = 16.373 (PEG tube = 1) - 16.976 (dementia = 1) - 1.278 (MMSE score) - 14.600 (decubitus = 1) + 37.619.

Table 4

Stepwise multiple regression analysis for duration of artificial nutrition

Variable	B	SE	ß	t value	p value
Constant	37.619	5.055		7.443	<0.001
Stomach tube	16.373	4.563	0.256	3.589	<0.001
Dementia	–16.976	4.901	–0.250	–3.464	0.001
MMSE	–1.278	0.482	–0.191	–2.652	0.009
Decubitus	–14.600	6.773	–0.155	–2.156	0.033

Stomach tube = Stomach tube (1) or not (0); dementia = dementia (1) or not (0); decubitus = (+) indicates 1, (-) indicates 0.

Discussion

This is the first study that evaluated both the cognition and functional ability of many patients receiving continuous artificial nutrition, as well as pneumonia, decubitus, and use of restraint [13]. Strikingly, in this study, most cases did not show the lowest score on the FIM motor domain. In the urination item of the FIM motor domain, only patients with a bladder catheter were scored 1 (the lowest score), and patients without a bladder catheter were scored 2 or higher [18]. We found that most patients (more than 80%) receiving artificial nutrition lived without a bladder catheter. Meanwhile, in the FIM cognitive domain and MMSE, most patients (70-80%) received the lowest score (5 points on the FIM cognitive domain or 0 points on the MMSE).

There have been three studies comparing survival time of patients with PEG and NG feeding [20,21,22]. Two studies reported a longer survival of patients with PEG than of those with an NG tube [21,22]. Moreover, Kumagai et al. [22] showed that patients receiving PEG feeding suffered less frequently from aspiration pneumonia than those with an NG tube. Meanwhile, Fay et al. [20] found no significant differences in survival times between patients receiving PEG feeding and those with an NG tube. In this study, we found a significant relationship between PEG use and longer survival of patients. Of course, it is possible that clinicians tend to adopt PEG for patients whom they expect to survive longer. Therefore, we cannot determine the cause-result relationship of PEG use with longer survival by regression analysis or a simple comparison study.

The prognosis of dementia patients with artificial nutrition has been controversial. It has been reported that dementia is a risk factor for shorter survival after PEG surgery [23,24]. However, in many recent reports, patients with dementia lived as long as those without dementia after PEG insertion [9,10,11,12]. In this study, dementia and cognitive test score are related to a shorter duration of artificial nutrition. Needless to say, the clinical characteristics of study samples affect the results greatly. We suppose that the difference in the results arises from the study population. In most previous studies, survival time of dementia patients was compared to that of patients with physical disease [9,10,11,12]. However, in this study, the survival duration of dementia patients was compared to that of patients with mental diseases. In this study, we found that dementia patients survived a shorter time than those with other diagnoses, which mainly consisted of psychiatric diseases. In the future, whether patients with mental diseases under artificial nutrition survive much longer than those with physical diseases under artificial nutrition should be clarified.

In this study, the functional state (FIM motor score) was not associated with the duration of artificial nutrition. On FIM motor scores, 28 patients (16.7%) scored 13 (lowest score), and 124 patients (73.7%) scored 14. Only 9.6% of the patients scored 15 or more. A so-called floor effect might obscure the effect of the functional state on survival. Of course, cognitive state is closely correlated with functional state in the severest stage of dementia. In this study, MMSE scores were associated with the duration of artificial nutrition. The patients with a longer duration of artificial nutrition showed lower MMSE scores. It is unlikely that patients with lower MMSE scores at the start of artificial nutrition live longer than those with high MMSE scores. Therefore, MMSE scores of patients with a longer duration of artificial nutrition decrease gradually in the course of artificial nutrition.

Serum hypoalbuminemia was reported to be a risk factor for survival in patients with PEG in many studies [3,9,10,11,25]. Regretfully, serum albumin was not evaluated in this study. However, hypoalbuminemia is reported to be a very important risk factor for decubitus in severe dementia [26]. Thus, decubitus in this study might be an index of poor nutrition. Our results are in line with those of Shah et al. [27].

There are several limitations to this study. First, it is a cross-sectional study, so prognostic factors such as survival time cannot be calculated. Second, in this study, only patients in psychiatric hospitals were included, and patients in nursing homes and general hospitals were not. Therefore, we cannot avoid a selection bias. However, this is the first study that identified the factors related to longer survival among inpatients, including those with mental disorders, with various methods of artificial nutrition.