Hypercaloric enteral nutrition in patients with amyotrophic lateral sclerosis: a randomised, double-blind, placebo-controlled phase 2 trial

Background— Amyotrophic Lateral Sclerosis


Background
6][7] For patients with feeding tubes, there are no nutritional guidelines nor is there consensus within the community regarding recommendations for enteral nutrition in ALS. 8 Multiple groups have reported an association between body mass index (BMI) and ALS survival, with BMI<18.5 associated with shorter survival [9][10][11] and moderate obesity (BMI 30-35) associated with slower disease progression and longer survival. 11,12 wo recent prospective studies found a reduction in ALS risk in persons who were overweight and obese. 13,14 dditionally, studies in the G86R and G93A superoxide dismutase 1 (SOD1) mouse model of ALS have shown that a calorie-dense high fat diet leads to weight gain and a delay in disease progression 15,16 while calorie restriction reduces survival. 17Given this epidemiologic and preclinical data, we hypothesized that a dietary intervention to increase body weight could improve survival in patients with ALS.We designed this phase II study to test whether hypercaloric diets with or without excess fat calories would be safe and tolerable in people with advanced ALS receiving enteral nutrition.

Study Design and Oversight
The High Fat/High Calorie Diet versus Optimal Nutrition in ALS clinical trial was an investigator-initiated, phase II, prospective, double-blind, placebo-controlled, randomized, multicenter clinical trial.The protocol and consent forms were approved by the Partners Institutional Review Board (IRB) and the IRB of 12 participating ALS centers, including the five clinics which form the Muscular Dystrophy Association (MDA) Clinical Trial Network (see complete list of participating sites, Appendix 1).The study was listed on clinicaltrials.gov(NCT00983983).The full protocol for this trial and statistical plan along with the supporting CONSORT checklist for trials of nonpharmacologic treatments are available as supplementary materials.All of the authors vouch for the fidelity of the study to the protocol.E.M. and A.M.W. had access to all the data and analyzed the data.A.M.W. was responsible for the decision to submit the manuscript.

Role of the funding source
The study was funded by the MDA with additional support from the National Center for Research Resources, the National Institutes of Health, and the Harvard NeuroDiscovery Center.Dietary supplements were purchased at cost from Abbott Pharmaceuticals who provided no financial support for the study.The sponsors had no role in the study design, data collection, analysis, interpretation, manuscript writing, or decision to submit.

Participants
From December 2009 to November 2012, adults with ALS receiving percutaneous enteral nutrition were recruited from ten of the twelve participating ALS centers.Recruitment was slower than anticipated due to the advanced nature of the disease, difficulty tolerating prestudy tube feeds, dependence on care-givers for transportation and help with feeding, fear of weight gain, limited projected survival, and eligibility for hospice services.In addition to advertising the study to providers and patients with the help of the MDA, the ALS Association, and the Northeast ALS Consortium, the home care agency Apria Healthcare sent mailings to 425 ALS patients receiving enteral nutrition, however this did not increase enrollment.
All participants provided informed consent prior to screening procedures.In the rare cases in which participants were unable to physically sign the informed consent, verbal consent was given in the presence of a witness.At screening, eligible participants had to have a diagnosis of suspected, possible, probable, or definite ALS by El Escorial criteria, 18 and had to have already tolerated enteral nutrition (for full Inclusion/Exclusion criteria, see Supplementary Table 1).Four people were ineligible based upon the exclusion criteria, which included requiring non-invasive ventilatory support for more than 10 hours/day, diabetes, myocardial infarction, or stroke (Figure 1, Enrollment and Outcomes).Cholesterol-lowering medications were prohibited during the study.

Randomization and Masking
Participants were randomized 1:1:1 to one of three dietary interventions: calorie replacement using an isocaloric diet with a goal of weight stability (control or Cntl), vs. a highcarbohydrate, hypercaloric diet with a goal of modest weight gain (high-carbohydrate/ hypercaloric or HC/HC), vs. a hypercaloric diet high in fat calories with a goal of modest weight gain (high-fat/hypercaloric or HF/HC).The randomization schedule was developed by the Biostatistics Center at Massachusetts General Hospital (MGH) with stratification by site in permuted blocks of three.Site investigators and coordinators were not aware of the block size.The Department of Metabolism & Nutrition Research Bionutritionist at MGH Clinical Research Center re-labeled tube feed cans prior to delivery to sites to maintain the blind.All coordination center staff, participants, investigators, site coordinators and site evaluators were blinded to treatment group assignment throughout the study.

Study Procedures
Energy requirements were estimated using indirect calorimetry adjusted for physical activity to define two levels of nutritional support: a replacement calorie (control) group, and two hypercaloric intervention groups (with or without modified fat content.At each in-person visit, the measured resting energy expenditure (MREE) was obtained using VMAX Encore metabolic carts (SensorMedics, Yorba Linda, CA) or equivalent to perform indirect calorimetry after 12 hours of fasting.Prior to randomization, normal caloric intake was calculated by nutrient analysis of four-day food records using Nutrition Data Systems for Research (NDS-R) version 2009. 19Physical activity coefficients were estimated using the Bouchard Three-Day Physical Activity Log. 20,21 stimated energy requirements were calculated by multiplying the physical activity coefficient times the MREE, or based on the daily intake required to maintain weight between the Screening and Baseline visits (a study window of 12-21 days), whichever was greater.Participants randomized to the control diet were prescribed Jevity 1.0 (Abbott Laboratories, Abbott Park, IL) to replace 100% of their estimated energy requirements.

Interventions
Because the high-fat diet used in the pre-clinical SOD1 mouse studies was 23% more calorically dense than the control diet, 16 we designed our two intervention arms to receive approximately 125% times the estimated energy requirements using either Jevity 1.5 or Oxepa (both Abbott Laboratories) with a goal of gaining approximately 0•5 kg/week.Oxepa contains 55% calories from fat compared to 29% in the Jevity products, including the omega-3 fatty acids eicosapentaenoic acid and gamma-linolenic acid.The percent of calories from protein (17%) was the same across all formula.If any participant experienced weight loss despite > 90% compliance with their tube feed formula, the Coordination Center Bionutritionist had the ability to increase their prescribed diet in a blinded manner.Because some participants were still able to consume food by mouth, the number of calories consumed could exceed the prescribed diets.The coordination center bionutritionist provided frequent phone consultations with each site to ensure protocol compliance and to encourage uniformity in nutritional management.A web-based electronic data capture system (Pharmaengine) was used to record all study data.The study intervention was designed to be four months long because the average survival time after PEG insertion is 4•5-8 months. 22Study activities were completed during five in-person visits over 4•5-5 months with a final follow-up phone call one month after discontinuing study diet.At the end of the four months on study diet, surviving participants resumed their pre-study diets or were prescribed new diets by their care providers.Participants were not informed of the total prescribed calories during or after the study in order to prevent unblinding.If participants agreed to long-term follow-up, vital status was verified at the end of the study.

Safety End Points
The primary outcomes of the study were safety and tolerability.Adverse events (AE) and serious adverse events (SAE) were coded using the Common Terminology Criteria for Adverse Events (CTCAE) version 3.0.Interim safety analyses were performed every six months during the course of the study by an independent data and safety monitoring board.

Tolerability
Participants kept daily diaries of their tube feed intake as well as any additional foods that they consumed by mouth.Participants were defined as tolerating the diet if they initiated study diet and were compliant with more than 80% of the prescribed study diet during the four months of the intervention.Participants were considered intolerant if they initiated study diet but failed to complete month four of the study on the originally assigned treatment for any reason.

Secondary Metabolic Outcomes
MREE and weight were measured at every in-person visit and participants were instructed to measure their weight at home on a weekly basis.Fat mass and lean body mass (LBM) were measured using Dual-energy X-ray Absorptiometry (DXA) using a Hologic Discovery A (Hologic Inc, Bedford, MA) or equivalent at baseline and month four at sites which had access to a DXA machine.Serum was collected at in-person visits after an overnight fast, frozen at −80°C, and sent to the MGH Research Laboratory for measurement of fasting lipids, beta hydroxybutyrate, albumin, pre-albumin, insulin, leptin, interleukin-6 (IL-6), a nd high sensitivity C-reactive protein (hs-CRP).

Secondary Efficacy Outcomes
Due to the small sample size, efficacy measures were included only as exploratory end points.Forced vital capacity (FVC) was collected at every in-person visit and the ALS Functional Rating Scale-Revised (ALSFRS-R) was collected at baseline and month four in order to reduce subject burden.

Statistical Analysis
Safety and tolerability analyses were performed on all participants who initiated study diets.Variables were summarized as frequencies and percentages, means and standard deviations, or medians and inter-quartile ranges as appropriate.Categorical and continuous variables were compared at baseline by Fisher's exact and Kruskal-Wallis tests, respectively.Survival was analyzed using intention-to-treat analysis, including all enrolled participants.Survival curves were estimated by Kaplan-Meier product-limit methods and compared by log-rank test.Survival time was defined as time to the earlier of death or tracheotomy with censoring at the end of planned follow-up or date lost to follow-up among surviving participants regardless of compliance with assigned study diet.Given the small sample size, significance of the log-rank test was confirmed by permutation test.Change over time of continuous, longitudinal measures was analyzed using random-slopes models.This analysis is unbiased if loss to follow-up was non-informative and data missing due to mortality were predictable from observed trajectories given assumptions of the model.Given the small number of participants, no adjusted analyses were performed.As an exploratory measure, not included in our original analysis plan, we performed the recently described combined assessment of function and survival (CAFS) using methods previously described. 23Two-sided p-values less than 0•05 were considered statistically significant without correction for multiple comparisons.All analyses were performed using SAS version 9•3 (SAS Institute, Cary N.C.).
The target sample size for this study was originally 60 subjects; however due to slow enrollment, we revised our target to 30 subjects or 10 per study diet.With this revised sample size, the study had 89% power to detect adverse events expected to occur among 20% of participants on a given study diet.

Study Population
A total of 28 people with ALS were screened and 24 participants were enrolled between December, 2009 and November, 2012.Seven participants were randomized to the control diet, nine to the HC/HC diet, and eight to the HF/HC diet.Four participants (one control, one HC/HC, and two HF/HC) withdrew after randomization but before starting study diet due to the burden of participation.The median number of participants per site was two (range one-six) and there was no imbalance in treatment allocation across sites (Figure 1. Enrollment and Outcomes).Baseline demographics and disease characteristics were wellbalanced across the three study arms (Table 1) except that participants in the HF/HC group had a higher ALSFRS-R (p=0•04).The controls had more males, fewer patients with bulbar onset, and higher LDL/HDL cholesterol levels (reported to be associated with improved survival in ALS). 24The HC/HC group had a lower baseline BMI, were slightly younger, had a shorter time since diagnosis, and had fewer participants who had been prescribed BIPAP, however these differences were not significant.

Safety and Tolerability Outcomes
Control participants were more likely to discontinue the study diet due to adverse events (three Cntl vs. zero HC/HC vs. one HF/HC) and less likely to complete the study on the intervention diet (17% vs. 88%, vs. 83%; p=0•03 and 0•08 for the difference in tolerability between Cntls and the HC/HC and HF/HC groups, respectively).
Participants on the HC/HC diet experienced fewer adverse events (AEs) and serious adverse events (SAEs) compared to control and HF/HC participants (24 AEs and 0 SAEs in HC/HC vs. 42 AEs and 9 SAEs in controls, vs. 49 AEs and 3 SAEs in the HF/HC arm, Table 2).The most common AEs were gastrointestinal (50% of HC/HC vs. 100% of Cntl vs. 100% of HF/ HC).None of the participants in the two hypercaloric arms experienced elevated serum bicarbonate (compared to three participants in the control arm), as we had postulated might occur due to respiratory weakness. 25In addition, there were no cardiovascular AEs or SAEs in the hypercaloric arms and the HF/HC diet was not associated with increased cholesterol or hs-CRP levels (Table 3).Finally, there was no evidence that the hypercaloric diets led to diabetes based on fasting blood glucose levels and serum insulin levels (Table 3).

Secondary Metabolic Outcomes
Participants randomized to the control arm were essentially weight-stable gaining on average 0•11 kg/month (95% CI −0•64, 0•86), although there was substantial variation in weight in the control arm (see Supplementary Figure 2A).On average, control participants consumed 1•21±0•26 times their estimated energy requirements, including both prescribed enteral nutrition and oral intake.

Secondary Efficacy Outcomes
There were no deaths in the HC/HC arm compared to three deaths in the control arm during the five-month follow-up schedule (log-rank test for survival p=0•03; Figure 2).There was one death in the HF/HC arm (log-rank p=0•23 compared to controls).While all of the deaths occurred within 30 days of stopping study diet, all deaths were due to respiratory failure and no deaths were considered related to study diet.Long term survival, recorded at the end of the study (up to 19 months after completing scheduled follow-up) is shown in Supplementary Figure 1.ALSFRS-R scores declined more slowly in the HC/HC arm −1•06 (95% CI −1•71, −0•41) points/month vs. −2•17 (95% CI −3•25, −1•1) in the surviving controls vs. −1•59 (95% CI −2•44, −0•74) in the HF/HC group, although this did not reach significance (p=0•07 for the comparison between the HC/HC and control groups, Supplementary Figure 2C).As the ALSFRS-R analysis excluded participants who died during the study because only a single follow-up assessment was completed, we also performed an exploratory post-hoc CAFS analysis.The mean (SD) joint rank scores were 14•8 (12•9) in the HC/HC arm vs. 6•0 (12•1) in the control arm vs. 9•3 (12•1) in the HF/HC arm (higher is better).The difference between the HC/HC and control arm was significant (Kruskal-Wallis p=0•01) while the difference between the HF/HC and control arm was not (p=0•26).There was no difference in the rate of decline in FVC (Table 3, Supplementary Figure 2D).

Discussion
ALS patients randomized to a high-carbohydrate hypercaloric formula with a goal of weight gain were less likely to experience serious adverse events including death during the study, compared to participants on a calorie replacement diet with a goal of weight stability.The study results support hypercaloric enteral nutrition as a novel and potentially robust nonpharmacologic intervention for this fatal disease.These results should be interpreted with caution given the small sample size.Another imitation which prevents the results from being generalizable is the fact that study participants had advanced disease and were malnourished, having lost almost 20% of their body weight on average by the time of enrollment.While this was a small pilot study, the results are consistent with epidemiologic and preclinical data showing that obesity in humans with ALS and hypercaloric diets in a mouse model of ALS are associated with improved survival.In addition to the safety data, the tolerability and secondary efficacy outcomes all favored the high carbohydrate hypercaloric arm.Possible risks of weight gain (hypercarbia, diabetes, or increased vascular events) were not observed during the study.
Our DXA results support the hypothesis that weight loss due to starvation may exacerbate muscle loss due to denervation by causing muscle catabolism. 2 Also consistent with prior reports, participants experienced an increase in body fat even while losing LBM. 8 Unfortunately, we did not have complete DXA data to compare the rates of LBM loss among the intervention groups.
We had initially hypothesized that the enteral formula containing 55% calories from fat (HF/HC) would be most effective at causing weight gain, as it most closely mimicked the diets used in animal studies.A recent study found that oral supplements containing a modest 35% calories from fat were slightly more effective at causing weight gain than supplements containing 0% calories from fat. 26 In our study, the HF/HC diet resulted in weight loss, despite participants consuming up to 174% of their estimated energy requirements.This may have been due to the lower tolerability and higher rate of gastrointestinal side effects compared to the HC/HC diet.Prior studies using Oxepa have not reported weight loss as a side effect. 27 were surprised by the number of calories required by participants to achieve weight gain.One possible explanation is that the observed MREE were inappropriately low; however, the MREE correlated well with the Harris-Benedict calculated REE and were measured by multiple machines at multiple sites.An alternative explanation is that selfreported activity levels were low; however, the estimated physical activity coefficients (mean 1•1, range 1•0-1•48) were consistent with prior doubly-labeled water experiments showing that advanced ALS patients' total daily energy expenditure is only 1•0-1•.2times their MREE. 28,29 nother explanation, supported by elevated hs-CRP levels in the majority of participants, could be that some participants were resistant to weight gain due to circulating cytokines as in cancer cachexia. 30Finally, the high frequency of gastrointestinal side effects in the control and HF/HC arms suggests that continued weight loss may have been due to gastrointestinal malabsorption.Our study had several strengths: first, by using enteral nutritional supplements, we were able to carefully control the number of calories

Systematic Review
We performed a systematic search of PubMed and Ovid for all clinical trials, observational studies, and reviews published between Jan 1, 1963 and February 7, 2014 in English, Spanish, Italian, German, and French.Search terms were nutrition supplements, hypercaloric, hyperalimentation, enteral nutrition, weight gain, oral supplements dietary supplements, and amyotrophic lateral sclerosis or ALS,

Interpretation
Currently the only FDA-approved therapy for ALS is riluzole, which provides only a modest survival benefit.Two small prior studies of oral supplements have shown some efficacy in causing weight gain in ALS, although these studies were underpowered to detect differences in disease outcomes in participants randomized to oral supplements.Our pilot study provides important preliminary evidence to support the use of hypercaloric nutritional interventions as a novel, low-risk therapy for the treatment of ALS.

Change over time in secondary and metabolic outcome measures, by treatment group
Parameter estimates for the interaction between treatment and time are shown (95% confidence intervals in parentheses) for BMI= body mass index; FVC: functional vital capacity percent of predicted; LDL= lowdensity lipoprotein; HDL= high-density lipoprotein; hs-CRP= high sensitivity C-reactive protein.Individual pair-wise p-values were calculated from contrasts using fixed effect solutions from models that included all three treatment groups.

Lancet.
Author manuscript; available in PMC 2015 June 14.NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1 .
Figure 1.Consort flow diagram of enrollment and outcomes.

Figure 2 .
Figure 2.Kaplan-Meier curves for overall survival.HC/HC= high-carbohydrate/hypercaloric diet; HF/HC= high-fat/hypercaloric diet; Cntl= control diet.The log-rank test for the difference across all treatments was p=0•06.The log-rank test for the difference in survival between the HC/HC and control arms was p=0•03.