NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Ferluga ED, Archer KR, Sathe NA, et al. Interventions for Feeding and Nutrition in Cerebral Palsy [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2013 Mar. (Comparative Effectiveness Reviews, No. 94.)
Key Findings and Strength of Evidence
State of the Literature
Feeding and nutrition problems are common among children with cerebral palsy (CP), and have significant health implications. Some patients with oral-pharyngeal dysphagia and gastroesophageal reflux (GER), particularly those with severe CP, are also at risk for recurrent aspiration which can lead to chronic pulmonary disease. Patients with feeding difficulties range from those with self-feeding skills to populations with severe disability (i.e., Gross Motor Function Classification System [GMFCS] V) who require extensive use of assisted technology and are dependent on others to feed them. Indeed, chronic pulmonary disease related to aspiration is a leading cause of death among patients with severe CP.27-30
This review is intended to gather together what data exist to support families and clinicians in making decisions about caring for children with CP as these decisions pertain to addressing problems with feeding and nutrition. It may also point the way forward to a future research agenda. Clinicians have available to them both behavioral and surgical treatments, which may be offered in sequence depending on the severity of the feeding issues, the effectiveness of initial treatment approaches, or new conditions brought on by prior treatment in the case of reflux that develops as a result of gastrostomy.
Ultimately, very few data exist to guide care. Our analysis of the behavioral literature consists of a summary of a good quality systematic review published in 2011, updated with one new case series evaluating a caregiver training program that is not manualized (documented in a manual so that it can be replicated). The surgical literature consists of a total of eleven studies meeting our criteria; studies are largely case series. One prospective cohort study was focused primarily on harms.
Across all interventions, the study populations are almost exclusively children with severe CP; when it is assessed, populations generally meet criteria for level IV or V of the GMFCS. Although study populations are generally assessed on overall severity (e.g., GMFCS) and weight, the use of other measures for growth and nutrition, and explicit characterization of the feeding challenges in the study population is lacking. Surgical outcomes data are available for fewer than 200 children and only one cohort study provides comparative data comparing surgical with oral interventions for any population of CP. Of note, those studies that do provide data on weight gain do so against reference populations of typically developing children. These are likely not appropriate reference standards; improvement in z-scores among children with CP may very well be clinically meaningful even if these children do not approach weight standards for the reference group.
Cerebral palsy is a group of conditions with wide variation in their expression thus it is difficult to assess whether outcomes observed in these studies represent likely, population-level effects. Furthermore, none of the studies providing effectiveness data was comparative. Interventions that address feeding and nutrition are often combined (e.g., positioning with pureed food, gastrostomy tube with fundoplication and supplementation, etc.), making it difficult to determine the independent effect of each.
Summary of Outcomes
Behavioral Interventions
We concur with the assessment of the author of the included systematic review that data on the effectiveness of sensorimotor interventions ( “techniques specific to the enhancement of oral-motor control aim[ing] to decrease or increase tone and inhibit abnormal reflexes that interfere with safe feeding”34) are conflicting, with the highest quality study (a good randomized controlled trial [RCT] of oral appliances) demonstrating no effect on feeding efficiency, but a fair quality RCT reporting positive outcomes in the subset of children with a history of aspiration using pureed foods. Other noncomparative studies demonstrated improvements in feeding skills, but these were poorly conducted case series.
No good quality studies were identified on the role of positioning to improve efficiency and safety measures, including reducing leakage and decreasing feeding time as well as reductions in choking. Despite the high risk of bias associated with the study designs, however, results were consistent that identifying an optimal position for the head and neck may help with feeding outcomes with individual studies reporting improvements in aspiration, feeding skills, feeding time, or food leakage
One study of limited quality reported that children with CP without speech, had better outcomes when eating mashed rather than solid food. Evidence for the benefit of oral appliances is mixed, with some studies reporting positive results in very short term outcomes, but few studies reporting on the outcomes of interest for our review. A fair quality study that included longer term data found no benefit of the appliances for feeding skills or weight gain beyond natural maturation of the children. Overall, there are few studies of any particular behavioral modality and those that have been conducted are small and lack rigor.
Surgical Interventions
We conducted a de novo review of the surgical literature. Data were available for two of the four surgical Key Questions—namely, Key Questions 3a and 3b.
The first question for which data were available was on the effectiveness of tube feeding versus oral feeding for children with CP without significant reflux. Evidence for the effectiveness of tube feeding (either g-tube or j-tube) comes from six case series87-94 and one prospective cohort study,95 designed to study the potential for overfeeding, and described below. All six case series focused on severely impaired children, and all reported significant increases in weight after gastrostomy, over six to more than 12 months. The most comprehensive case series reported improvements on all weight and growth related outcomes (weight, head growth, linear growth, arm circumference and skinfold thickness), including closing the gap significantly with a normally developing reference population, and significantly more than would have been expected without intervention. Followup continued to 12 months post surgery, with data available on 46 of the initial 57 children, and 6 unavailable due to loss to followup. The five other case series with growth data also reported significant pre-post increases in weight, but data on other measures were unreported or inconsistent. One case series also assessed health care utilization as a proxy for overall health and found the number of hospitalizations significantly reduced over the year following gastrostomy.94
Two studies that reported on quality of life (QOL) measures89,93 in one, parental QOL improved significantly, in tandem with decreases in feeding time, after gastrostomy. In the other, most parents (98%) expected that their child's QOL would improve with gastrostomy, but did not report that to be the case after surgery.
No studies directly compared the use of g-tube with fundoplication with oral feeding for the treatment of reflux. One RCT compared two forms of plication (fundoplication versus vertical gastric plication),99 and in one case series children undergoing gastrostomy plus Nissen fundoplication showed improvements in reflux symptoms and weight gain but not episodes of pneumonia; 30 percent of participants had recurrent reflux post-surgery. In the RCT, the Nissen fundoplication group had a greater decrease in the total number of reflux episodes, percentage of pH, and longer reflux episodes, whereas the vertical gastric plication group showed only a change in the pH percentage.
Harms associated with tube feeding and reported in the comparative literature include surgical harms, infection, increased rates of reflux requiring further treatment, and potential overfeeding. The harms reported with gastrostomy are not significant. Overall, rates of peritonitis were low, ranging from 2 to 5 percent, but one study reported minor site infections at 59 percent and leakage at 30 percent. In the studies focused on fundoplication, major complications were experienced by 14.3 percent of the children in each arm of the RCT, and minor harms included aspiration, urinary tract infection and pneumonia. In the case series early complications (within one week of fundoplication) included pneumothorax, stoma ulcers, and cellulitis, all in patients undergoing laparoscopic fundoplication. Late complications (i.e., more than one week post-fundoplication) included bloating, diarrhea (dumping syndrome), and intestinal obstruction; these events were all reported in individuals undergoing open fundoplication.100
Prior, retrospective studies, have suggested that gastrostomy may be associated with higher than expected rates of mortality. In the surgical studies included in this review, reported death rates ranged from 7 to 29 percent, with varying follow up times. The investigators report that these rates are comparable with those in other studies of feeding tubes, but there are no data to suggest whether these are reduced relative to what would have occurred absent treatment. In one study that did not meet our inclusion criteria, 2 of 15 patients with cerebral palsy and no gastrostomy died, relative to 17 of 47 with gastrostomy. This was a retrospective review of cases, however, so any comparison would likely be confounded by indication, as only children with severe feeding and growth problems are typically treated surgically.101 In another retrospective study that did not meet our inclusion criteria, survival rates after gastrostomy or jejunostomy were 83 percent after 2 years and 75 percent after 7 years.102 It is unclear what these apparently high rates of mortality mean relative to expected mortality, and to families facing the reality of a severely undernourished or growth deficient child.
Three studies were specifically intended to analyze harms of tube feeding: one on the potential for tube feeding to induce reflux96 and two on the potential to overfeed.41,95 Two found positive associations with the harmful outcome(overfeeding or reflux).95,96 The clinical importance of these and other harms, relative to the potential danger of not intervening is unclear and likely must be assessed in the context of each individual patient and family. Clearly, surgical interventions can lead to increased weight gain; the degree to which harms outweigh those benefits likely depends on the starting point of the individual, family stressors, and the degree to which harms can be mitigated using appropriate feed and other approaches – an area that warrants continued research.
The frequent report of GER that develops after gastrostomy may be balanced by the study investigators' observations that it is frequently managed medically. In terms of overfeeding, one study on this subject demonstrates that tube fed children may be at risk for obesity without careful attention to the content and quantity of their food products; one study assessing the effects of a low energy feed, which increased weight without increasing fat mass, suggests that such formulas may have the potential to reduce the risk of overfeeding.41
No studies directly compared the use of g-tube with fundoplication with oral feeding for the treatment of reflux. One RCT compared two forms of plication (fundoplication versus vertical gastric plication),99 and in one case series children undergoing gastrostomy plus Nissen fundoplication showed improvements in reflux symptoms and weight gain but not episodes of pneumonia; 100 30 percent of participants had recurrent reflux post-surgery. In the RCT, the Nissen fundoplication group had a greater decrease in the total number of reflux episodes, percentage of pH, and longer reflux episodes, whereas the vertical gastric plication group showed only a change in the pH percentage. Major complications were experienced by 14.3 percent of the children in each arm of the RCT, and minor harms included aspiration, urinary tract infection and pneumonia. In the case series early complications (within one week of fundoplication) included pneumothorax, stoma ulcers, and cellulitis, all in patients undergoing laparoscopic fundoplication. Late complications (i.e., more than one week post-fundoplication) included bloating, diarrhea (dumping syndrome), and intestinal obstruction; these events were all reported in individuals undergoing open fundoplication.100
Strength of the Evidence for Effectiveness of Therapies
Overview
We used the included systematic review on behavioral interventions to assess strength of evidence, translating the assessment used in that review into levels used in the EPC program. Behavioral studies including in the prior review34 were small, typically short-term, and typically conducted using pre-post designs subject to bias. The author of the systematic review used a modified Sackett approach (Table 5) to assess the strength of the body of evidence. We have translated those assessments into EPC program equivalents in Table 14.
Table 14
Strength of the evidence for behavioral interventions assessed in Snider review.
Strength of evidence for behavioral interventions ranges from low to moderate. The moderate rating for the positive effects of oral appliances on sensorimotor outcomes is based on one good and one fair quality RCT and additional supporting studies of varying designs. Effects on eating efficiency and swallowing were not consistent, and the small sample sizes suggest imprecision. The low strength of evidence for beneficial effects of positioning, altering food consistency, and feeding devices on all outcomes is due to the lack of RCTs and generally small sample sizes. Studies typically reported some positive effects on mealtime length and eating efficiency; however, rigorously conducted studies are lacking. The strength of the evidence for the effects of oral sensorimotor interventions and oral appliance on feeding safety and efficiency is insufficient based on a paucity of rigorous studies.
Longer term studies are lacking across all interventions; thus, the durability of effects is not clear. Studies also did not consistently assess harms, though aspiration and swallowing difficulties, which may be related to the underlying condition as well as the intervention, are reported in some. Overall, more data on greater numbers of participants, including adults as well as children, are needed to understand the effectiveness of behavioral approaches.
We also assessed strength of evidence for six primary outcomes associated with feeding tubes in comparison with oral feeding, and for feeding tubes with fundoplication to address reflux: changes in growth outcomes, respiratory outcomes including reflux, quality of life, long term morbidity and mortality, and harms. We found the evidence to be insufficient to low for all outcomes (Tables 15–16).
Table 15
Outcome, strength of evidence domains, and strength of evidence for feeding tubes (KQ3a).
Table 16
Outcome, strength of evidence domains, and strength of evidence for fundoplication (KQ3b).
The low strength of evidence for the effects of gastrostomy on increasing growth measures, including weight, is based on a clearly significant effect measured in five case series and one prospective cohort study and in a small number of children. Additional data are needed on greater numbers of children to better quantify expected effects, particularly in subgroups by severity and age, and to better understand the implications of observed harms. Long term effects are unknown as data on mortality are short term only. Nonetheless, it is clear that, in children with significant feeding difficulties, most of whom present significantly underweight, tube feeding leads to weight gain. Evidence is currently insufficient to assess whether and to what degree fundoplication is effective specifically to treat children with CP who present with significant reflux.
Findings in Relationship to What Is Already Known
The findings from this review for both surgical and nonsurgical interventions are consistent with both prior systematic reviews and relevant clinical practice reviews, as there is little robust evidence about the effectiveness of either behavioral or surgical interventions for children with feeding difficulties and CP. As a result, there are no definitive clinical guidelines for physicians due to this uncertainty and it is difficult for parents to make an informed decision about the risks and benefits of these interventions, particularly surgical interventions. Findings from this review and all prior systematic reviews call for well-designed randomized controlled trials of both medical management and surgical interventions to assess for short and long term outcomes, including harms, for both children and adults.
Our findings support the conclusions of the prior systematic review on behavioral interventions, including oral sensorimotor, oral appliances, and positioning. Furthermore, due to the recentness and completeness of that review, our review identified only one additional study, a case series. The case series, at high risk of bias, addressed caregiver training.
Regarding surgical interventions (gastrostomy), our review found no randomized trials or high quality observational studies comparing gastrostomy with oral feeding for patients with CP and feeding difficulty. Only five case studies and one retrospective cohort study informed this question. Our findings are consistent with the prior 2004 Cochrane systematic review which also found no randomized clinical trials comparing g-tube with oral feeding for patients with CP and feeding difficulties, and another systematic review which included observational studies, found severe methodologic limitations to the studies, and therefore insufficient evidence regarding the effects of gastrostomy compared with oral feeding.
Harms associated with feeding interventions have not been thoroughly reviewed in prior systematic reviews, and observational studies continue to raise questions about the risks and benefits of surgical interventions for children with severe CP and feeding difficulties. This review continues to find limited data on harms.
Applicability
Applicability of Studies of Behavioral Interventions
Studies of behavioral interventions to date have been limited in scope and focus on a limited selection of outcomes of interest. Studies typically provided limited data on heath outcomes including hospitalizations, antibiotic use, patient and family satisfaction and quality of life, measures of family stress, pain/comfort. In addition to the recent systematic review from Snider and colleagues, we located one case series based in the home among child-caregivers pairs in Bangladesh and assess applicability in Table 17. The study focused on caregiver training related to diet, food consistency, appropriate utensils, and postural and physical support for positioning and feeding. Evidence from this study is likely primarily applicable to younger children who are able to eat at least some foods orally. The approach studied may not closely match interventions available in practice as it was conducted in the home setting, which is likely highly variable, and was not well-described. Thus, individuals wishing to infer the potential results of clinical practice based on the available research need to assess carefully the degree to which the study methods matched those available and used in practice. Ultimately, the effectiveness of behavioral interventions within and outside of this limited sample and setting is currently unknown.
Table 17
Applicability of studies of behavioral interventions.
Applicability of Studies of Surgical Interventions
All of the studies of surgical interventions focused, appropriately, on severely impaired individuals, generally GMFCS levels of IV or V (Table 18). Those studies that provided data to characterize the participants indicated that children in the studies had experienced substantial lack of growth, for up to 12 months prior to intervention. Participants were followed for 6 months to over a year. Studies assessed outcomes of interest to clinicians and caregivers of individuals with CP including changes in measures of growth, hospitalizations, and chest infections. The two studies of fundoplication for reflux similarly included children, but their level of functional impairment was not clearly described. Studies were not designed to assess subsets of individuals as defined by types of feeding disorders or specific surgical intervention.
Table 18
Applicability of studies of surgical interventions.
Implications for Clinical and Policy Decisionmaking
The effectiveness of feeding and nutrition interventions for individuals with cerebral palsy remains largely unknown with strength of evidence not exceeding moderate for any intervention. Nonetheless, clinical decision makers can use this review to understand what interventions are available, what outcomes have been seen, and to some degree, to balance potential harms. When a child has a severe feeding disorder and is unable to consume adequate nutrition and is affected by frequent aspiration and pneumonias, the health outcomes can be dire. Understandably, treatment decisions must be made, even with inadequate evidence. Parents and providers contemplating gastrostomy can use the review to help understand potential effects on their quality of life and that of the child, potential harms that may occur, and potential tradeoffs related to social functioning. They should do so in light of the severity and other issues facing the individual child and family. Of note, nonclinical considerations may include family stress and pressures related to providing optimal care for the individual child. Stressors associated with caring for a severely disabled child and the potential impact of feeding interventions on the relationship with the child should not be underestimated and may play into decisionmaking along with the limited clinical evidence available. Ideally, this review will help policymakers and researchers understand what types of studies are essential to lead to more informed clinical decisionmaking.
Limitations of the Comparative Review Process
This review was focused on identifying and assessing effectiveness literature. As such, we used stringent requirements for the presence of appropriate comparison groups for the behavioral and nutrition literature and for pre-post data for the surgical literature. Thus, we excluded some literature lacking pre-intervention data that may have provided some information, although it is unclear that these studies would have contributed to an assessment of effectiveness. In addition, we required that at least 80 percent of the population have cerebral palsy. It is possible that some literature on populations with feeding disorders for reasons other than cerebral palsy may have included data that would be relevant. However, it is not clinically straightforward to determine which individuals or groups with feeding disorders are similar enough in etiology and presentation that their data would be appropriate for our population of interest.
Limitations of the Evidence Base
Behavioral studies are uniformly small, and usually underpowered to demonstrate clinical effectiveness. Outcomes tended to be short-term feeding skills (e.g., drooling, chewing), and followup was typically less than 6 months post-intervention. The presence of one good systematic review provides a comprehensive overview of the state of the literature. Across the board, rigorous, comparative (ideally RCTs) studies should be done of behavioral interventions. While RCTs are best for establishing causal inference, it is likely that they may not be optimal study designs for all questions that are important in this field of study. In particular, eliminating the confounding effect of potential mediators and moderators could result in not fully understanding the complexity in the natural history and appropriate treatment of feeding challenges. A range of study designs will be necessary to address the breadth of important questions currently unanswered.
The most significant challenge in this body of surgical literature is the lack of comparative data and, absent direct comparisons, lack of understanding of the natural history of feeding disorders among children with CP that might be useful in weighing the risks associated with surgery against potential benefits. The field generally considers comparative studies to be unethical, in part because no nonsurgical approaches to care have been clearly shown to be effective to serve as comparison treatments. Children presenting for surgical intervention are generally (as shown in all studies) substantially underweight and demonstrate additional deficiencies in nutrition. To delay treatment is ethically challenging. Denying or delaying nutritional treatments including food thickeners or special formulas similarly poses ethical challenges. Larger, well characterized series may be the only reasonable solution to obtaining good outcomes data. It is possible that a registry could be useful to capture detailed data on patients in addition to consistent outcomes data. Adequately capturing patient characteristics, including type of CP and functional level, is imperative for increasing our understanding of the risks and benefits of therapeutic approaches.
Another fundamental problem with the current studies is that they are relatively small; it is not possible to assess effectiveness of treatment approaches in subsets of individuals characterized by severity, specific feeding challenges, presence of reflux and type of procedure. Additionally, multiple interventions may be frequently used in these populations, particularly those with severe CP. Certainly, additional or continued multicenter series are needed. The studies included in this review were fairly short term and constrained by the requirement that they provide data both before and after surgery. Thus, harms were typically limited to those that occurred within a year of surgery; followup in surgical studies ranged widely from roughly two months96 to over 18 years in one study.90 Retrospective studies not meeting the criteria for inclusion in this review have been used to suggest that tube feeding is associated with increased mortality. One of the studies included in this review94 was intended to prospectively address this question; this work should be continued and extended. Additional prospective data with potential confounders clearly characterized is necessary to better understand whether the mortality rates observed in these studies are due to the surgery or the cerebral palsy and associated respiratory disease.
Research Gaps and Areas for Future Research
The study of feeding and nutritional interventions for individuals with cerebral palsy is a nascent field, but certainly one that is growing. Rigorous, comparative studies of behavioral interventions need to be conducted; good RCTs are largely missing from the literature. Nonetheless, current research is available to provide potential directions for study. For example, studies of sensorimotor interventions currently provide conflicting evidence and more rigorous evidence is needed to answer the open question as to whether they can be effective at improving outcomes. Studies of positioning are also warranted. Studies should also compare each of the behavioral interventions with one another, with extensive characterization of the participants to better understand what works for which patients. Research should also investigate promising approaches used in other populations with similar impairments for potential applicability to individuals with CP. Similarly, research should address nutritional interventions such as food thickeners or vitamin supplementation.
Foundational research is needed to establish the most appropriate, patient-centered outcomes that are important to families of individuals with CP. Along these lines, qualitative and mixed methods approaches may be useful for understanding the experiences, preferences, needs, and strengths of families and caregivers. The degree to which improved changes are considered target outcomes by families is not well established. It is also not clear whether short-term outcomes translate to longer term health outcomes. We note that there is a complete lack of studies designed or powered to identify modifiers of effectiveness of the behavioral interventions.
As noted above, the ethics of conducting comparative surgical studies or studies of nutritional interventions in the absence of appropriate comparison groups may preclude rigorous comparative designs. Case series can be conducted in ways that move them closer to providing effectiveness data; in addition, well developed registries may provide a source of data for observational study designs. Of particular importance is the need to conduct large enough studies to fully characterize both participants and interventions so that the question of whether treatment approaches are better for individuals who, for example aspirate or do not aspirate, can be answered. Patients with cerebral palsy are heterogeneous in many ways, including severity and comorbid conditions; rigorous subgroup analyses are needed to obtain data for targeting treatment. Furthermore, they and their families already experience substantial burden in terms of healthcare and other stressors. Recruitment and retention is likely to be a challenge, and may be a reason for the relatively poor evidence base to date.
In all types of interventions, data are absent on the role of feeding interventions for adults with CP. In addition the interventions included in this review, the importance of the nutritional make-up (energy composition) of the food products themselves are necessary. Prospective, comparative studies should be carefully conducted to determine what type of nutrition is appropriate for obtaining positive health outcomes without inducing excessive weight gain.
Considerable uncertainty remains concerning harms over both the short and long term. Harms associated with feeding interventions have not been thoroughly reviewed in prior systematic reviews, and observational studies continue to raise questions about the risks or benefits of surgical interventions for children with severe CP and feeding difficulties.
Conclusions
Evidence for behavioral interventions for feeding disorders in cerebral palsy is insufficient to moderate. Some studies suggest that interventions such as oral appliances (moderate strength of evidence for effects on oral sensorimotor skills) may be beneficial, but there is a clear need for rigorous, comparative studies. Evidence for surgical interventions also is insufficient to low. All studies to date demonstrate significant weight gain with gastrostomy. Results for other growth measures are mixed, and substantial numbers of children remained underweight, although given a lack of appropriate reference standards for the CP population, these results should be interpreted cautiously. Considerable uncertainty remains concerning harms over both the short and long term. Harms with gastrostomy can be common, and include overfeeding, site infection, stomach ulcer, and reflux. Mortality rates range from 7 to 29 percent. Longer term, comprehensive case series are needed to understand potential harms in the context of benefits and potential risks of not treating.
- Discussion - Interventions for Feeding and Nutrition in Cerebral PalsyDiscussion - Interventions for Feeding and Nutrition in Cerebral Palsy
Your browsing activity is empty.
Activity recording is turned off.
See more...