The present study revealed that NGF is not associated with a decreased pneumonia risk in older patients with dementia who are cared for with in-home healthcare system. Meanwhile, hospitalization rate and one-year mortality were similar between AHF and NGF. However, all measured outcomes showed a similar, or rather less significant trend of increased risk in the NGF group. This study could provide preliminary evidence for tube feeding in home healthcare studies. Our study showed that the risk of pneumonia diagnosis was not significantly lower but revealed a non-significant trend of increasing risk of pneumonia in the NGF group. In older patients with dementia, gradual impairment of swallowing develops with disease progression. Owing to the loss of a normal clearance mechanism, the most frequent consequence is aspiration pneumonia, which typically occurs during eating when pharyngeal materials enter the lower airway, followed by infection with nonpathogenic flora (18).
A review of 19 cohorts suggested that tube feeding could not reduce the risk of aspiration pneumonia (19). Subsequently, a study compared the incidence of aspiration pneumonia in individuals with NGF, gastrostomy, jejunostomy, and oral feeding. The authors found that after a 6-month follow-up, more patients (58%) in the tube feeding group developed aspiration pneumonia than in the oral feeding group (54% in NGF, 67% in the gastrostomy group, 75% in the jejunostomy group, and 17% in the oral feeding group). Furthermore, the study demonstrated a significant difference in risk of aspiration pneumonia between oral feeding and tube feeding (p < 0.01) (20). Importantly, a recent study indicated that tube feeding was associated with around two-fold increased incidence of aspiration pneumonia diagnosis (RR = 2.32; 95% CI, 1.22–4.40) (21). To be more specific, oral secretions and regurgitated gastric contents are severe causative factors for aspiration pneumonia, and feeding tubes provide no protection efficacy (19). As opposed to stroke, dementia is a progressively deteriorating disease (22). Individual strategies are required to manage different pathogeneses. Tube feeding is generally reserved for short-term (< 1 month), reversible, or unconscious individuals, such as those requiring post-surgery support and discharge. However, if the enteral tube feeding is anticipated more than 1 month, percutaneous endoscopic gastrostomy (PEG) is preferred due to less treatment failure, fewer complications and better nutritional status (23) (24) although one-year pneumonia risk showed no significant difference between the NGF group and the PEG group (aOR = 0.51; 95% CI, 0.18-1.47) (25). For mild or moderate dementia, PEG placement should be preferred over the long- term placement of nasogastric tube, which was apparently the standard care in current guideline (26). Of note, the initiation of tube feeding in patients with severe dementia is not recommended (26). Additionally, pressure sores were potential predictive factors for pneumonia in the study participants. Proper positioning and regular position changing are useful strategies for the care of pressure sores (27). Given the limited data available currently, position changing may be beneficial in pneumonia care (28).
With respect to hospitalization rate and duration, we found no significant difference in our study. A prior prospective observational study with 6-month follow-up also detected no difference in hospital admission between tube feeding and oral feeding (21). Hospitalization leads to distress and is associated with poor outcomes in advanced dementia patients. However, older patients with dementia are frequent visitors to hospitals for several reasons. A prospective cohort study of 617 patients investigated the prevalence and mortality in an acute care setting revealed an increasing admission rate with age (29.6% for patients aged 70–79years and 75% for patients aged >90 years). Urinary tract infection or pneumonia was the primary cause. Moreover, these patients showed a higher mortality rate (adjusted mortality risk 4.02, 95% CI, 2.24–7.36) (29). Most hospitalizations among older patients with dementia are avoidable. Additionally, invasive treatments in acute hospital care may not meet the needs of this group and lead to injuries. Even infectious diseases such as pneumonia could be treated effectively in a nursing home rather than in the hospital (30). In the present study, the major cause of hospital admission was pneumonia. Increased Hb levels may be a protective factor for decreasing hospitalization rates and duration. Importantly, Hb could be an early indicator of the nutritional status of hospitalized older patients (31). The home care system in Taiwan made great efforts to advocate the avoidance of frequent admissions and provided phone consultation as required (13).
The present retrospective observational study of home care residents showed no difference in mortality rate between AHF and NGF in patients with advanced dementia patients. However, this study revealed a non-significant trend of increased mortality rate in NGF group. A previous study suggested that 85.8% patients with advanced dementia (stage 7 on the Global Deterioration Scale) suffered from eating problems, with the 6-month mortality rate of 38.6% (32). Furthermore, a longitudinal study in Taiwan investigated the survival outcome of patients with dementia. The authors concluded that 77% patients who were placed with a nasogastric tube died within 6 months. Therefore, nasogastric tube placement was recognized as a risk factor for increased 6-month mortality (33). A 2009 Cochrane systemic review, which included seven observational studies, revealed insufficient evidence of increased survival rate in tube feeding groups (9). A large prospective database study provided valuable evidence in survival analysis. They evaluated 36,429 nursing home residents with severe dementia and eating difficulty. In the study, 1956 residents were under PEG insertion compared with 34,536 residents who were fed orally. They found no significant difference of survival between both groups (34). In line with earlier reports, our study did not favor tube feeding for decreasing the mortality rate. In previous studies, tube feeding group was mainly composed of PEG feeding. However, in this study, we added valuable evidence comparing AHG and NGF in mortality rate analysis. All above evidences may indicate that, once dementia progresses to a severe stage, with the development of eating problems, the mortality rate increases regardless of placement of feeding tube. In addition, increased one-year mortality rate was observed in subjects with low Norton scale in our study. Similar to prior studies, patients with lower Norton scale score were associated with higher mortality rate in older patients (35) (36).
Despite significant efforts to improve oral intake, most patients with dementia develop eating difficulties during the severe stage. A nasogastric tube was considered a convenient and efficient approach for individuals with feeding problems. However, NGF frequently presented with complications (e.g., blockage, dislodgement, pneumonia, trauma from insertion, and use of physical or chemical restraints) (37) (38). When comparing NG and gastrostomy groups, a significantly higher number of tube-related complications were reported in the NG group (aOR = 0.18; 95% CI, 0.05–0.65) (25). AHF in patients with advanced dementia should be considered an alternative to tube feeding (39). Management of eating difficulties in advanced dementia remains controversial. The goals should emphasize on avoiding functional decline and adopting safe and effective oral feeding. A comprehensive swallowing assessment could help identify dysphagia etiology and offer adequate strategies for improving swallowing function. Furthermore, several approaches can increase the efficiency of careful hand feeding (e.g., modifying the feeding position, feeding skills, and changing both the feeding environment and texture of food). Of note, feeding skills are essential for preventing aspiration (e.g., the chin tuck posture and head rotation toward the weak side). Patients with an impaired oropharyngeal musculature can more easily control thickened food than thin fluids. Modifying the food texture of solids or liquids increases feeding effectiveness, particularly in patients with severe dementia (40). Nevertheless, making the modified food acceptable and appetizing is vital to increase compliance. Furthermore, an adequate environment setting (e.g., removing distractions and scheduling mealtime with family) could make patients enjoy their social time, as opposed to isolated tube feeding. In the present study, well-trained nurses provided all these useful strategies for oral feeding of residents at the home health care system. To date, tube feeding still plays a role because of individuals’ concerns over the discomfort of being thirsty or hungry. However, the consequences of forgoing tube feeding (e.g., discomfort or pain) have not been investigated thus far. A prospective observational study of 178 nursing home patients attempted to address this issue. They observed different items creating a discomfort index. Shortness of breath, restlessness, observation of pain, and dehydration were associated with increased discomfort. Given the evidence that discomfort levels were not associated with forgoing tube feeding (41), additional efforts could be placed on evaluating the patients’ quality of life (42).
Our study expanded the study group from nursing home to home healthcare in the community settings, which is one of the significant care systems for severe dementia in Taiwan. Furthermore, relatively large number of patients receiving NGF is a useful comparison with other Asian countries, where the prevalence of NGF is high.
In hospitals or nursing homes, well-equipped facilities and specialized health care provide a high quality of care. Meanwhile, there was only one or no caregiver in the home care system, with inadequate training. Results from the home care system are essential for community practice. We believe that this preliminary study of tube feeding issues in subjects with severe dementia requiring home healthcare in Taiwan will help healthcare providers, patients, and families by providing a more realistic understanding of nutritional support in this population.
The present study has several limitations. First, for ethical reasons, conducting randomized controlled trials would be extremely challenging. In this retrospective study, the recruited subjects were already receiving NGF or AHF, and were not randomly enrolled. Therefore, the baseline characteristics of both groups were different. We restricted the study group only to complete bedridden status and severe dementia to make a balance in the study. Although we took the effort to select more comparable subjects in the NGF and AHF groups, the patients belonging to the NGF group were more compromised, as shown by the lower Barthel index and albumin level. Furthermore, intervention (NGF) and control (AHF) groups were not well matched, and the difference of sample size between these two groups was not negligible. However, this could also explain the high preference of NGF in patients with severe dementia because of its convenience and efficiency. Moreover, if outcome data from multiple centers could be collected, the study groups could be widely applied in general home healthcare population. Finally, no data were available on the following aspects: quality of life, objective assessment of discomfort, pain scale, use of physical and chemical restraints, and physical function. Quality of life is more important than life span, and it is usually neglected.