Feeding Dementia Patients via Percutaneous Endoscopic Gastrostomy

Despite such findings, feeding by PEG tube remains a highly common method of nutritional support to keep patients alive for months or even years. In fact, according to a 2003 analysis by Mendiratta and colleagues,¹ PEG tube use increased by 38% in hospitalized elderly patients, and 1 in 10 PEG tubes was placed in a patient with dementia. In this article, we comment only on the clinical evidence for the use of feeding tubes in patients with dementia. The ethical, moral, and religious implications of these findings are beyond the scope of this discussion.

Discussion

There is much conflicting evidence about whether tube feeding improves clinical outcomes. In a significant review published in 1999, Finucane and colleagues² surveyed the medical literature between 1966 and 1999 and came to the conclusion that “for severely demented patients the practice [of tube feeding] should be discouraged on clinical grounds.” They found no data supporting clinical improvement as a result of tube feeding in these patients. In 2009, Finucane’s paper was described by DeLegge³ as “pivotal,” defined as one that “changes the
direction of a practice in medicine or creates a breakthrough discovery.” The medical literature subsequent to Finucane’s paper supports DeLegge’s characterization. Reviews and editorials^4,5 strongly discourage the use of PEG tubes in patients with dementia, basing their conclusions in large part on the review by Finucane and colleagues.² For example, Casarette and colleagues⁵ make the categorical statement that artificial  nutrition and hydration “does not improve survival rate among patients with dementia.”

Shortly after Finucane’s review, Callahan and colleagues⁶ reported detailed findings in a prospective cohort study involving 150 elderly patients in whom a PEG tube had been placed. These were seriously ill patients; 24 patients died even before the baseline assessment and 22 more patients died before the 60-day follow-up. Of the 72 remaining patients, at least 70% had no significant improvement in functional, nutritional, or subjective health status. Furthermore, many patients experienced symptomatic problems attributable to tube feeding.

In a commentary accompanying Callahan’s report, Finucane and Christmas⁷ responded: “Only a minority of patients showed any measurable improvement. Functional status, serum albumin, and weight were better at follow-up in very few patients.” What is not mentioned, however, is that 30% of the patients in Callahan and colleagues’ study did have an improvement in serum albumin and only 5% showed a decline. Weight was maintained in 98.5% of patients. It would seem that the goal of significant improvement in the various indices that were assessed in these seriously ill patients was an unreasonable one. Prevention of deterioration would have been a more logical expectation. The data seem to indicate that PEG placement can maintain life and select indices of physiologic function in a significant percentage of patients with close to end-stage disease. These were patients who could take no significant oral nutrition, and in the absence of some kind of assisted feeding, they would likely not have survived for more than a few weeks.

Finucane and Christmas⁷ advised physicians to tell families considering PEG the following: “We do know that if the PEG is placed, there is a very high chance that he/she will die soon.” While this statement was unquestionably true for the desperately ill patients included in Callahan and colleagues’ study, the premise for the decision regarding whether to place a PEG is misleading: the patients probably did not die because of the PEG, but rather with the PEG, and perhaps despite the PEG. While concerns of a poor prognosis despite PEG placement for persons with severe dementia are probably valid, there has been a widespread extension to dementia in general, without clear evidence to justify this. This extension would seem to be unwarranted because of the paucity of data regarding PEG placement during the earlier stages of dementia, before severe malnutrition and deterioration have occurred.

In contrast, several reports in the literature support the nutritional benefits and life-prolonging potential of PEG placement in adults with dementia. Lindemann and Nikolaus⁸ report that more than half of their patients with dementia lived longer than 6 months with a PEG tube, and their data suggest that one should consider placing the PEG tube at an earlier stage in the dementia. In another study, Shapiro and Friedmann⁹ reported their experience treating dementia patients in Jerusalem-based nursing homes, noting that 43% of their patients were alive at 2 years after PEG placement and 31% were still alive at 3 years after PEG placement. A wide variety of institutions throughout Israel have had similar experiences with thousands of patients; more extensive use of PEG tube feeding is found in Israel than in other Western countries for a variety of cultural reasons.^10,11

In 2008, a group of investigators in Japan examined the survival rates after PEG placement in dementia patients and cognitively intact patients.¹² The study found similar rates of survival at 12 months in both groups, with 51% of dementia patients and 49% of cognitively intact patients being alive. In addition, 20% of patients with dementia lived longer than 3 years; however, these investigators also reported increased mortality if a patient’s serum albumin level was low and if significant comorbidities were present. The authors also summarized the data for all 10 studies in the literature that reported survival benefits, and a substantial percentage of the patients in these studies survived beyond 1 year.¹²

When taking into account the pros and cons of PEG placement based on the evidence in the literature, there are several key considerations that emerge. First, no patient can survive without adequate nutrition. When one removes a feeding tube without substituting another route of nutrition, death becomes imminent. PEG tube withdrawal has become an acceptable protocol in many Western countries to end the life of patients who are in a permanent vegetative state. Not feeding a dementia patient leads to the same outcome. An article by Hurley and Volicer¹³ about end-of-life care describes in a positive light the removal of a PEG tube in a female patient with advanced Alzheimer’s disease with the intent of causing death. PEG placement had sustained this patient’s life for more than 3 years. The patient’s daughters had the tube removed because they became dismayed by their mother’s quality of life, and she died shortly thereafter.

Second, although PEG tubes can serve as a convenient way to provide sustenance and extend life, oral feeding is preferred whenever possible. In dementia patients, oral feeding may be very costly in time and effort and is therefore often impractical in many nursing homes; however, it should be encouraged. Geriatric patients and chronically ill patients are frequently malnourished, and this condition is often not diagnosed at its early stages. There should be ongoing evaluation of dementia patients for the earliest signs of undernutrition. Oral feeding is obviously the ideal route to maintain adequate nutrition whenever possible, but at the earliest signs of malnutrition, alternate methods of feeding should be considered before more serious debilitating conditions develop, which may render PEG placement futile. We have previously made these recommendations in the medical literature.¹⁴

A third consideration is that complications associated with PEG placement are likely to be more frequent and more serious in patients with advanced diseases. If the decision to use parenteral feeding is delayed until the serum albumin level drops too low, wound healing will likely be impaired and infections will be more likely to occur. This is why we support close monitoring of nutrition status in patients with dementia, even in the early stages when self-feeding is still possible. When malnutrition is discovered early, it can be addressed while the risk of complications is lower and outcomes are more likely to be improved.

Finally, until there is more substantial evidence on the potential benefits and harms of PEG tube feeding in elderly dementia patients, the decision to use or end this kind of long-term nutritional support in end-of-life care appears to belong to patients and their family caregivers in keeping with their personal and cultural preferences.

Based on the currently available literature, it is clear that more research on when and for whom to implement PEG feeding is needed, particularly in the setting of dementia care. In a 2009 Cochrane review, Sampson and colleagues¹⁵ found that no randomized controlled trials and only seven observational controlled studies have been conducted to evaluate outcomes of enteral tube feeding for older people with advanced dementia who develop problems with eating and swallowing and/or have poor nutritional intake. None of these seven studies showed conclusive evidence of prolonging survival or improving quality of life, and none fully explored adverse effects of enteral feeding in this specific patient population. The researchers urged, “This area is difficult to research but better designed studies are required to provide more robust evidence.” We strongly endorse this recommendation, especially the need for randomized controlled trials, before concluding whether PEG tube feeding in dementia patients can and should be used to prolong life. Specifically, we suggest that patients who cannot be adequately nourished through oral feedings be divided prospectively and randomly into two grouops: one to be fed by PEG and the other not. One might then learn whether there are specific subgroups of patients with dementia in whom the balance of benefit/risk favors use of PEG and perhaps others who reap no benefits.

Conclusion

Several seminal papers reporting unfavorable results following PEG placement in patients with advanced dementia have led to the widespread conclusion that the use of enteral feedings in patients with dementia is not indicated. But other reports and experiences raise questions about these categoric conclusions. We urge clinicians to take a more nuanced
position until prospective randomized controlled studies to differentiate between subgroups of patients with dementia with varying benefit/risk ratios are undertaken to shed more conclusive light on this issue.