What is the clinical effectiveness of NG feeding tubes compared with PEG for patients with HN cancer?

Four SRs^1,9,18,19 were identified. No additional published RCTs beyond those included in the SRs were identified although one RCT in progress was found.¹⁵ Table 2 presents overviews of the four SRs and Appendix 2 provides greater SR detail. The lead authors of the SRs were from Canada,¹⁹ the Netherlands,¹⁸ the UK¹ and China.⁹ Two SRs limited their analyses to RCTs which were rare, with one meeting inclusion criteria in the Cochrane review¹ and two different trials in the SR from the Netherlands.¹⁸ The other two SRs included less rigorous evidence, primarily cohort studies (seven in one and eight in the other).^9,19 Patient groups varied from very specific (e.g., one SR included only studies on adults with stage III or IV HN SCC receiving curative CT plus RT¹⁹) to quite general (e.g., the most recent SR from China included studies on adults with HN cancer receiving CT and / or RT⁹). This may explain the variation in numbers and types of included studies in the four SRs.

Table 2

Overview of Included SRs.

Detail about an RCT in progress:

NIHR-NETSCC, the “TUBE trial” (ISRCTN48569216 in “Controlled Clinical Trials”):¹⁵ A three-center study in England will examine whether a definitive RCT is feasible in patients with Stage III and IV HN SCC undergoing CT and RT. The objective is to compare prophylactic PEG (or percutaneous fluoroscopic gastrostomy [PFG]) versus oral feeding plus as-needed NG tube feeding. It is hoped that about 60 patients will be randomized and the trial will run from June to December 2014. Randomization will be administered centrally by the Newcastle Clinical Trials Unit internet-accessed secure web-based system. Due to the obvious visual differences in the technologies being studied, blinding of patients and providers is not possible. Funding is from the HTA programme of the National Institute for Health Research. The study protocol was published May 1, 2014 and the estimated date of study publication is July 2016. Planned outcomes for this feasibility RCT are to:

-

Assess willingness to be randomized

-

Assess retention and drop-out rates from each arm

-

Refine interventions and study processes to inform definitive trial design

-

Estimate parameters to inform sample size for a definitive trial

What are the evidence-based guidelines regarding the use of NG feeding tubes in an outpatient setting for patients with HN cancer?

Four CPGs were identified that provided enough methodological detail to assure the reader that fairly rigorous processes had been followed (Table 3).^20–23 The four CPGs originated from Australia;²³ Ontario, Canada;²¹ the United Kingdom;²⁰ and the United States²² and publication years spanned 2009 to 2014. Processes for the very recent United States National Comprehensive Cancer Network (NCCN) were not completely clear in the 190-page CPG but the guidance was included here as another author described NCCN cancer CPGs as “the most comprehensive and widely used oncology standard in clinical practice in the world…recommendations are now accepted by the Centers for Medicare and Medicaid Services and most private insurance companies.” (p.187)²⁴

Table 3

Overview of Included CPGs.

What is the clinical effectiveness of NG feeding tubes compared with PEG for patients with head or neck cancer?

The AMSTAR quality assessment tool for SRs¹⁶ was used to assess SR quality and scores were assigned (Table 6 in Appendix 2). All SRs were based on extensive and well documented literature searches. All SRs employed at least two independent data extractors. Despite these efforts, the evidence base is limited as there have been very few RCTs. For example, one RCT of 33 randomized patients²⁵ was included in the recent Cochrane review,¹ the RCT being switched to a prospective non-randomized study as few patients would consider randomization¹² and two others were included in an SR from the Netherlands (although these were considered by the Cochrane reviewers and ultimately did not meet their inclusion criteria).¹⁸ The other two SRs included more studies but these were of less rigorous designs.^9,19 Of these two SRs, the earlier review¹⁹ commented that the available studies were not of high quality but did not report using a tool to assign scores, whereas the more recent SR⁹ used the Newcastle-Ottawa scale to assess study quality and assigned scores of 5 to 9 with a mean value of 6.5 (median 6) out of a maximum score of nine.

What are the evidence-based guidelines regarding the use of NG feeding tubes in an outpatient setting for patients with HN cancer?

The AGREE II instrument for CPG quality¹⁷ was used as a general guide with particular attention paid to CPG scope; CPG methodology, i.e., extent and reporting of the literature search, types of included evidence, grading of evidence and recommendations; and developer and funder potential conflicts-of-interest (COI). See Table 4 for CPG strengths and limitations.

Table 4

Strengths and Limitations of Included CPGs.

Cancer Care Ontario (CCO) CPG²¹ developers sought the highest quality CPGs available from other groups upon which to base their work. They determined that a CPG from the Scottish Intercollegiate Guidelines Network (SIGN) best addressed questions about clinical management and a CPG from the National Institute of Health & Care Excellence (NICE) in the United Kingdom best addressed questions about organization of care. These two base CPGs were assessed by three independent reviewers for quality, currency, content, consistency, and acceptability/applicability using the Appraisal of Guidelines Research and Evaluation (AGREE) instrument. Information was updated via a Medline literature review to late 2007 (details are provided). This material was used to develop 177 recommendations. A 40-person working group from seven professions involved in the care of patients with HN cancer reviewed the recommendations via a modified Delphi survey process. Recommendations were marked according to degree of consensus (i.e., complete or not).

The ENT UK CPG document²⁰ is extensive and was developed by the input of members of 10 professions; 128 experts are credited for their contributions. Unfortunately, there is little detail about the actual methodology of CPG development including how literature was chosen. Recommendations were graded using the SIGN system of A to D based on a grading system of the evidence ranging from 1++ (high quality meta-analyses, SRs of RCTs, or RCTs with a very low risk of bias) to 4 (expert opinion). The methods used to develop the CPG are not included in the document.

A CPG developed by the Cancer Council of Australia in 2014²³ includes a detailed description of the methodology. The Cochrane Database of Systematic Reviews, CENTRAL, MEDLINE, EMBASE, CINAHL and AMED databases were searched in December 2009 with the search repeated in January 2011 (limited to English language). The search terms are provided. Two researchers reviewed the literature to select included materials and the strength of the evidence was assessed (Levels I to IV) using the level of evidence rating system recommended by Australia’s National Health and Medical Research Council (NHMRC). Study quality was assessed via a tool supported by the American Dietetic Association that includes four relevance questions and ten validity questions to assess the strength of the study design. Recommendations were graded using the NHMRC Grades of Recommendation (A to D). The CPG is published on a wiki website and it is anticipated that it will continually be reviewed as new literature is published and based on input from other stakeholders.

Finally, the very recent CPG developed by the National Comprehensive Cancer Network (NCCN)²² is extensive and involved the input of many experts (predominantly physicians). The 190-page CPG focusses on many aspects of treating patients with HN cancer but there is no description of the methods used to develop the guidance. There are two pages of advice on principles of nutrition, supported by three references, but no assessment of the evidence or grading of evidence or recommendations.

What is the clinical effectiveness of NG feeding tubes compared with PEG for patients with head or neck cancer?

Four SRs were identified that included from one to eight primary studies.^1,9,18,19 One of the SRs was a recently updated Cochrane review¹ that included one study – a small (n=33) RCT published in 2009. Another SR also limited itself to RCTs and included two studies. Individual review findings are reported in Table 7 of Appendix 2 and demonstrate that not one of the four SRs was able to conclude that the newer technology, PEG, was definitely superior to the traditional use of NG tube feeding employed on an as-needed basis. For example, the most recent SR⁹ included eight studies of less rigorous design and the authors were tentative in their conclusions. They noted that the evidence did not illustrate differences between groups in maintenance of weight or survival, though several differences between groups were noted (Table 7). From the perspective of PEG superiority: tube dislodgement was lower, PEG was more suitable for long-term feeding (> 30 days), and PEG allowed greater mobility, enhanced cosmesis and improved QOL. However, on the negative side, the included studies suggested that PEG is associated with delay in return to oral diet, prolonged duration of RT, increased incidence of pain, increased incidence of dysphagia and markedly increased cost.

What are the evidence-based guidelines regarding the use of NG feeding tubes in an outpatient setting for patients with head or neck cancer?

The CPGs were reviewed for advice on the use of NG tube feeding in outpatients with HN cancer (Table 5); however, recommendations were seldom specific to the outpatient versus the inpatient setting and also gave little guidance about use of NG tube versus PEG feeding (rather using a generic “tube feeding” description). Where there was specific detail about NG tube of PEG feeding, this is included in Table 5.

Table 5

Summary of CPG Content Related to Tube Feeding for Adults with HN Cancer.