A Comprehensive Examination of Percutaneous Endoscopic Gastrostomy and Its Association with Amyotrophic Lateral Sclerosis Patient Outcomes

Abstract

There is literature discord regarding the impact of percutaneous endoscopic gastrostomy (PEG), or "feeding tube", on amyotrophic lateral sclerosis (ALS) outcomes. We assess one of the largest retrospective ALS cohorts to date (278 PEG users, 679 non-users). Kruskal-Wallis and Kaplan-Meier analysis compared cohort medians and survival duration trends. A meta-analysis determined the aggregate associative effect of PEG on survival duration by combining primary results with 7 published studies. Primary results (p < 0.001) and meta-analysis (p < 0.05) showed PEG usage is associated with an overall significant increase in ALS survival duration, regardless of onset type. Percent predicted forced vital capacity (FVC %predict) ≥50 at PEG insertion significantly increases survival duration (p < 0.001); FVC %predict ≥60 has the largest associative benefit (+6.7 months, p < 0.05). Time elapsed from ALS onset until PEG placement is not predictive (p > 0.05). ALSFRS-R survey assessment illustrates PEG usage does not slow functional ALS pathology (p > 0.05), but does stabilize weight and/or body mass index (BMI) (p < 0.05). Observed clinical impression of mood (CIM), was not impacted by PEG usage (p > 0.05). Overall results support PEG as a palliative intervention for ALS patients with ≥50 FVC %predict at PEG insertion.

Keywords: ALS; dysphagia; malnutrition; motor neuron disease.

Figure 1

Differences in ALS survival duration based on ALS onset type and PEG usage. (A) All PEG users (n = 279) had a significantly (p < 0.05) longer disease duration than all non-users (n = 649). (B) Limb onset PEG users (n = 138) had significantly longer disease durations than bulbar onset PEG users (n = 133) (p < 0.05). (C) PEG is associated with a significant increase in survival duration in limb onset PEG users (n = 138) compared to limb onset non-users (n = 448) (p < 0.05). (D) PEG usage is associated with a significant increase in survival duration in bulbar onset PEG users (n = 133) compared to bulbar onset non-users (n = 164) (p < 0.05).

Figure 2

Association of percent of predicted forced vital capacity (e.g., FVC %predict) at the time of PEG placement date on ALS survival duration (in months). Patients with a percent predict of ≥60% (n = 95) had significantly higher disease duration than patients with percent predict of <50% (n = 57) (p < 0.05). Patients with a percent predict between 50 and 60 did not have a significantly different disease duration than the <50% group or the ≥60% group (p > 0.05 in both cases).

Figure 3

Effect of trimester of disease duration at PEG placement on remaining disease duration. Each of the PEG users’ disease durations was divided into three equal time periods, called trimesters. (A) Patients who began PEG in the first trimester (n = 116) did not have significantly different disease durations than those who began PEG in the second trimester (n = 87) (p > 0.05). (B) Patients who began PEG in the second trimester (n = 87) did not have significantly different disease durations than patients in the third trimester (n = 53) (p > 0.05). (C) Patients who began PEG in the first trimester (n = 116) did not have significantly different disease durations than those who began PEG in the third trimester (n = 53) (p > 0.05).

Figure 4

Effect of PEG Placement on change in patient scores for the Amyotrophic Lateral Sclerosis Functional Rating Scale Revised survey (ALSFRS-R). PEG user ALSFRS-R declines were divided between the first ALS clinic visit to PEG placement and from PEG placement to the last recorded clinic visit. ALSFRS-R declines were not significantly different after PEG placement (p > 0.05). Thus, PEG placement and usage did not impact rate of ALSFRS-R decline.

Figure 5

Kaplan–Meier survival curves. (A) PEG users (n = 279) compared to non-users (n = 49). (B) PEG users with FVC %predict <50 at initial PEG placement (n = 57) compared with FVC %predict ≥50 (n = 124) and FVC %predict ≥70 (n = 54). (C) Bulbar onset PEG users (n = 133) compared with bulbar onset non-users (n = 64). (D) PEG users who began using PEG in the first trimester of disease duration (n = 116) compared with those who began in the second (n = 87) and third trimesters (n = 53). (E) Limb onset (n = 138) PEG users compared with limb onset non-users (n = 448).

Figure 6

Comparison of median disease durations (with error bars representing interquartile range) of PEG users and non-users from the meta-analysis. Disease durations of PEG users and non-users were recorded from published studies that found either significant or insignificant increase in disease duration correlated with PEG usage (see Table 2). Prior study results were combined with the present study’s results in order to determine the aggregate effect of PEG usage on survival duration. Fisher’s test results (see Methods) illustrate a significant aggregate effect that results in a positive associative survival benefit to ALS PEG users. For consistency, the disease duration for the meta-analysis is defined as time elapsed since patient-reported first symptom onset until recorded date of death.