Scolaris Content Display Scolaris Content Display

Cochrane Database of Systematic Reviews Protocol - Intervention

Systemic nutritional interventions for treating foot ulcers in people with diabetes

This is not the most recent version

Collapse all Expand all

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

To evaluate the effects of systemic nutritional interventions on the healing of foot ulcers in people with diabetes, in people of any age, in any care setting, with either type 1 or type 2 diabetes.

Background

Description of the condition

Definition

A Diabetic foot ulcer (DFU) is a non‐healing or poorly healing, partial or full thickness wound below the ankle, in an individual with diabetes (Lavery 2008; Sanders 2011).

Etiology

Foot ulcers in people with diabetes are diagnosed by taking a history and a physical examination combined with monofilament testing for neuropathy and non‐invasive testing for arterial insufficiency (Armstrong 1998; Kruse 2006). People with diabetes may have either neuropathic, arterial or venous components to their ulcer, or a combination of all three (Nelzen 1993). Neuropathy is a disease affecting nerves, with the symptoms dependent on whether motor, sensory or autonomic nerves are affected (Marcovitch 2006). Long standing hyperglycaemia results in nerve damage associated with autonomic, sensory and motor neuropathy. All three categories of neuropathy can be found in people with diabetes: sensory neuropathy causes a loss of pain sensation; autonomic neuropathy can cause either anhydrosis (dry skin) or hyperhidrosis (excessive sweat), both of which affect skin quality; and motor neuropathy causes weakness of muscles and structural changes to the foot (Merriman 1995).

Peripheral vascular disease (PVD) is the narrowing of the vessels (arteries and veins) in the legs and is common in people with diabetes, with atherosclerosis caused by hypertension being the most common cause of PVD (Marcovitch 2006). Atherosclerosis is an accumulation of cholesterol plaques on the artery walls, or the degeneration of the artery walls caused by hypertension and high cholesterol (Marcovitch 2006; Tortora 2006). Bilateral macro‐ (large vessel) and micro‐ (small vessel) vascular insufficiency is common in people with diabetes due to the prevalence of atherosclerosis.

In people with diabetes, a combination of PVD and neuropathy dramatically increases the likelihood of the development of a DFU. Indeed, the United Kingdom Prospective Diabetes Survey (UK PDS 1998) found in their large, multicentre study of people with newly diagnosed type 2 diabetes that 10% had some level of neuropathy and vascular disease on diagnosis, however it was not stated as to whether this is common in the general population or is specifically related to those with diabetes.

Neuropathic DFUs can be caused by acute mechanical or thermal trauma or from repetitively or continuously applied mechanical stress, whereas ischaemic foot ulcers in people with diabetes are often caused by a single mechanical damage to poorly perfused (and often friable) tissues (Cavanagh 2005). Therefore ischaemic foot ulcers are often caused by a single trauma to poor quality skin and are slow to heal as a result of poor blood supply to the wound, whereas neuropathic DFUs are due to reoccurring pressure and shearing forces and, due to their lack of pain sensation, the person is often not aware of any potential problems until an ulcer has actually formed. Inappropriate footwear and walking barefoot or in sandals are common causes of DFUs (Boulton 2004).

Foot ulcers in people with diabetes can occur regardless of the type of diabetes: type 1 diabetes is caused by an absolute deficiency of insulin secretion; and type 2 is caused by a combination of resistance to insulin action and an inadequate compensatory insulin secretory response (ADA 2008). Several studies have found that the incidence of a foot ulcer is around 82% to 87% more common in people with type 2 diabetes than people with type 1 diabetes; (Jeffcoate 2006; Oyibo 2001b; Piaggesi 1998). Interestingly it has been noted that both type and duration of diabetes has no effect on ulcer outcomes (Abbott 1998; Oyibo 2001b).

Prevalence

Prevalence gives an indication of the total number of cases of disease present at any one time, and includes both old and new cases of the disease (Marcovitch 2006). Prevalence is important as it gives an overall picture of the current disease situation whereas incidence, the number of new cases that occur during a particular timeframe (Marcovitch 2006), is useful when planning for the future and when comparing and contrasting clinical outcomes and auditing care delivery.

Wild 2004 estimates that the global prevalence of diabetes in 2002 was 2.8% or 171 million persons, furthermore, this rate is projected to rise to 4.4% or 366 million persons by 2030. The projected rise in prevalence is based on demographic changes alone and thus is a conservative assumption as it does not include other risk factors such as obesity and reduced physical activity. The International Diabetes Federation (IDF 2012) released less conservative global prevalence estimates of 8.3% or 371,329 persons for the year 2012 and a projected estimate of 9.9% or 551,870 persons by 2030. From an individual country perspective, Diabetes UK estimate that the prevalence in the UK in 2006 was 3.5% and increased to 4.6% in 2012 (Diabetes UK 2006; Diabetes UK 2012). In Ireland, conservative estimates suggest that in 2008, 141,063 or 4.7% of adults had diabetes and in Ireland it is predicted that by 2015, 193,944 or 5.6% of adults will have diabetes (Balanda 2010). The prevalence of diabetes in Canada is also rising from 5.9% or 1.65 million in 2008 to 6.5% or 1.9 million in 2012 (Statistics Canada 2010; Statistics Canada 2013). In 2000, the prevalence of diabetes in Australia was 7.5% (Dunstan 2001). However, the Australian National Health Survey estimates more conservative figures; in 1995 it estimated 2.4% which rose to 3.8% in 2007‐2008 (ABS 2012). The American Diabetes Association (ADA) notes that in 2010, diabetes affected 25.8 million people or 8.3% of the US population, however these statistics are estimates as they include diagnosed and undiagnosed diabetes (ADA 2011). Unlike many countries, incidence and prevalence of diabetes in Scotland is collected on a national level. In 2011, the incidence of diabetes in Scotland was 258,570 or 4.9% of the population (NHS Scotland 2012). In 2011, the European prevalence rate was 8.1% or 52.8 million people (IDF 2012). The European region country with the highest rate of diabetes was Russia with 10% or 12.6 million people; in contrast, Moldova had the lowest rate of just 2.8% (IDF 2012). In 2011, the island nation of Kiribati was the country with the highest world wide diabetes prevalence of 25.7% (IDF 2012).

Diabetic peripheral neuropathy is one of the most common predisposing factors leading to foot ulceration; indeed, people with diabetes are 50 times more likely to develop a foot ulcer than their non‐diabetic counterparts (Tyrell 1999). The prevalence of DFUs in the diabetic population is suggested to range from 3% to 10% (Borssen 1990; NICE 2004; SIGN 2010). However, Singh 2005 argues that rates are much higher, suggesting that the life time risk of a person with diabetes developing a foot ulcer is as high as 25%. Recurrence rates of DFUs post‐healing are very high, at rates of 60.5% and 66% (Hunt 2009; Peters 2007). Furthermore, diabetes is the leading cause of non‐traumatic limb amputation in the world (Tyrell 1999). Within 18 months following amputation, almost 50% of people will develop a DFU on the other limb, and of these people, 58% have further amputations within three to five years. It is worthy of note that the three‐year mortality rate after the first amputation is between 20% and 50% (Boulton 2005; Mulder 2003).

Incidence

The ADA found that 1.9 million new cases of diabetes were diagnosed in people aged 20 years and older in 2010 (ADA 2011). The Scotish incidence of diabetes, across all ages, was 18,985 in 2012 (NHS Scotland 2012). The annual incidence of DFUs is 2.5% to 10.7% in resource‐rich countries (e.g. Australia, Finland and the US) (Hunt 2009). The incidence of lower limb amputation in Scotland was 1359 or 0.6%, however 13,091 or 5.1% of people with either type 1 or type 2 diabetes were recorded to have had a DFU (NHS Scotland 2012).

Health‐related quality of life

DFUs are associated with disability and reduction in health‐related quality of life (HRQoL) (Nabuurs‐Frassen 2005; Ragnarson Tennvall 2000; Siersma 2013). Indeed, the presence of a DFU has been shown to have an independent effect on HRQoL of both patients and carers. The largest impact on HRQoL for patients is related to mobility, with the largest impact on carers being emotional (Nabuurs‐Frassen 2005). Indeed, frequent hospital visits, the need for wound care, and the fear of amputation were among the factors responsible for the impact on carers' HRQoL. For the individual, the longer the DFU is present the greater the burden, with Nabuurs‐Frassen 2005 noting a moderate improvement in HRQoL once the DFU is healed and a further improvement three months post‐healing. Interestingly, Ragnarson Tennvall 2000 identified that patients who healed post‐major amputation have a lower HRQoL than those who heal without the need for an amputation, indicating that the method of healing has an influence on post‐DFU HRQoL. Thus, owing to the negative influences on HRQoL related to the presence of a DFU and its associated treatments, Siersma 2013 argues that to improve HRQoL, treatment should not be only focused on DFU healing, but rather on all aspects of quality of life. This emphasises the importance of adopting a multidisciplinary approach to the prevention and management of the DFU, in order that the needs of the person with diabetes are adequately addressed.

Cost

It is estimated that in 2001, DFUs, excluding amputations, cost the US healthcare system USD 10.9 billion (EUR 8.09 billion) (Gordois 2003a; Shearer 2003), whereas, in the UK, in the same year, Gordois 2003b estimated that GBP 252 million (EUR 300 million) was spent on DFUs. A large European, multicentre, prospective study, averaged the cost of treating a DFU at approximately EUR 10,000, however these costs vary between different outcome group, furthermore, the data were derived from seven countries including both low and high cost countries (Prompers 2008). In general, however, it is suggested that the costs attributed to the diabetic foot in the US and Europe are as high as 20% (7% to 20%) of total diabetes expenditure (Apelqvist 1994; Girod 2003; Gordois 2003a; Gordois 2003b; Harrington 2000; Saar 2005; Sedory‐Holzer 1998; Shearer 2003; Van Acker 2000; Van Houtum 1995; Vozar 1997).

Duration of DFU influences the cost of overall treatment, and Apelqvist 1994 noted that 54% of people with a DFU heal within two months, 19% heal in three to four months and 27% heal in more than five months. A more recent prospective study (Zimny 2002) found the average DFU duration was 133 days for ischaemic wounds, 123 days for neuroischaemic wounds and 77 days for neuropathic wounds. Furthermore, the costs rise markedly with DFU severity (Gordois 2003a). For example, a retrospective cost analysis (Sedory‐Holzer 1998) noted an incremental cost increase related to the progression of DFU from Wagner grade 1 to Wagner grade 5. Management of Wagner grade 1 or 2 averaged at USD 1929 (EUR 1426.25) compared to Wagner grade 5 which averaged at USD 15,792 (EUR 11,683.93). A more recent study (Stockl 2004) also found a correlation between costing's and Wagner grade where the cost of Wagner grade 1 averaged at USD 1892 (EUR 1399) and Wagner grade 5 averaged USD 27,721 (EUR 20,509). Furthermore, Reiber 1995 noted that the cost for a major amputation doubles that of a minor amputation USD 12,879 compared to USD 26,940.

Eighty per cent of people with diabetes live in low and middle income countries (IDF 2013). Amputation and death are common outcomes of ulceration in Tanzania where treatment is inaccessible or unavailable to many (Abbas 2005). The availability of diabetes treatments and care in developing countries is a broad subject but perhaps low tech approaches such as nutritional intervention could be of particular benefit in these settings.

Classification

DFU status is recorded and monitored using a classification system (Oyibo 2001a). There are several different systems in use currently, however none is universally accepted and different centres use different systems (Schaper 2004). Despite this, Oyibo 2001a highlights the Wagner Wound Classification (Wagner 1981) and the University of Texas (UT) Diabetic Wound Classification system (Lavery 1996) as two of the most widely used and well established classification systems available. Detailed scoring systems, such as the UT system, offer a valuable method for comparison of data from different diabetic foot centres, however simplistic scoring systems, such as the Wagner system, may be used in clinical practice (Karthikesalingam 2010).

The Wagner Wound Classification System

The Wagner system assesses ulcer depth and the presence of osteomyelitis (infection of bone (Marcovitch 2006)) or gangrene (the death or decay of body tissue caused by a cessation of the blood supply (Marcovitch 2006)).

The grades range from 0 to 5. Grade 0 indicates a pre‐ or post‐ulcerative lesion, grade 1 indicates a partial/full thickness ulcer, grade 2 indicates an ability to probe to tendon or capsule, grade 3 indicates a deep ulcer with osteomyelitis, grade 4 indicates a partial foot gangrene and grade 5 indicates a whole foot gangrene (Wagner 1981).

The University of Texas Diabetic Wound Classification System

The UT system assesses ulcer depth, the presence of wound infection and the presence of clinical signs of lower extremity ischaemia (lack of blood flow due to contraction, spasm, constriction or blocking of the arteries (Marcovitch 2006).

The UT system includes both grades and stages. The grades range from 0 to 3. Grade 0 indicates a pre‐ or post‐ulcerative site, grade 1 indicates a superficial ulcer not involving tendon, capsule or bone, grade 2 indicates an ulcer penetrating to bone or joint, and grade 3 indicates an ulcer penetrating the bone or joint. The stages range from A to D. Stage A is a clean ulcer, stage B is a non‐ischaemic infected ulcer, stage C is an ischaemic non‐infected ulcer, and stage D is an ischaemic infected ulcer. Each ulcer is classified in the UT system with both a grade and a stage combined (Lavery 1996).

Pillars of diabetic wound healing

HbA1c is an index of the average blood glucose concentration over approximately the preceding six weeks (Kumar 2005). Control of HbA1c is paramount in wound healing, and poor metabolic control has long been associated with non–healing wounds (Boyko 1999; Greenhalgh 2005). In addition to HbA1c control offloading, debridement (the mechanical removal of foreign material and damaged tissue from a wound (Marcovitch 2006)) and infection control are seen as the corner stones to foot ulcer treatment in people with diabetes. Other supplementary interventions are also used in the treatment of DFUs such as hyperbaric oxygen therapy (Abidia 2003; Faglia 1996), growth factors (Tsang 2003) and tissue‐engineered skin (Marston 2003). The role of nutrition in the treatment of DFUs has largely been under researched, as the focus of nutrition has tended to be on achieving tight control of HbA1c.

Description of the intervention

Nutritional status is a dynamic entity reflecting physiological requirements, nutritional intake, body composition and function (BDA 2001). The presence of a wound impacts on nutritional status due to the metabolic cost of repairing tissue damage, sepsis (poisoning by the products of the growth of micro‐organisms in the body (Marcovitch 2006)) and nutrient losses via wound exudate (a fluid that has exuded out of a tissue or its capillaries due to injury or inflammation (BDA 2001; Marcovitch 2006)). Thus, it is reasonable to assume that the nutritional status of the DFU patient may interfere with the healing process (Bowling 2004; Tatti 2012). Nutritional interventions should be commenced when a nutritional assessment has determined that nutritional status is sub optimal (BDA 2001). Indeed, Leininger 2002 argues that assessing a patient's nutritional status and providing adequate support when deficiencies are suspected is a fundamental component in the management of individuals with wounds.

Methods to improve or maintain nutritional intake are known as nutritional support (NICE 2006), however, nutritional intervention, nutritional support, nutritional treatment and medical nutrition therapy are all interchangable terms for systematically attempting to improve a persons nutritional status. Categorisation of nutritional supports differ between NICE 2006, Bowling 2004 and BDA 2001, however, oral supplements, enteral (the introduction of nutrients directly into the gastrointestinal tract by nasogastric tube (Marcovitch 2006)) and parenteral (the introduction of nutrients by intravenous administration (Marcovitch 2006)) are the three common categories.

Oral nutritional support

NICE 2006 lists fortified food, additional snacks and sip feeds as methods of oral nutritional support (ONS). Some ONS products are nutritionally complete and can be taken to supplement the diet, or as a sole source of nutrition, however others only contain certain nutrients and are designed to supplement the diet (Bowling 2004). Standard ONS includes polymeric‐, peptide‐, or amino acid‐based (types of protein) supplements and also those where novel substrates have been added, such as glutamine (an amino acid synthesised within the body from glutamic acid and used in preventing immunosuppression after exercise and as an aid in recovery after a critical illness (Marcovitch 2006)), fish oils, arginine or antioxidants. We will include nutrient‐based novel substrate in this review. We will exclude prebiotics, probiotics and synbiotics as they are not nutrients. Schrezenmeir 2001 describes prebiotics, probiotics and synbiotics as non‐digestible food ingredients.

Enteral

Enteral tube feeding is the delivery of a nutritionally complete feed via a tube into the stomach, duodenum or jejunum (NICE 2006). NICE 2006 recommend enteral tube feeding for those requiring long‐term (four weeks or more) nutritional support.

Parenteral

Parenteral nutrition is the method of providing nutritional support to an individual whose gastrointestinal tract is not functioning or is inaccessible (BDA 2001). Nutrients are delivered directly into the circulatory system via a dedicated peripherally inserted central catheter (BDA 2001). NICE 2006 suggests that micronutrients, trace elements, electrolytes and nutrients can be added to parenteral feeds to meet the patient's individual needs. Parenteral nutrition is considered a short‐term method of nutritional support (less than 14 days), however in very special considerations it may be used as a long‐term method (NICE 2006).

NICE 2006 advises the use of oral, enteral and parenteral nutrition alone, or in combination, for people who are either malnourished or at risk of malnutrition. However, Gottschlich 2001 argues that enteral nutrition is superior to parenteral nutrition, particularly during the early phase of wound healing.

How the intervention might work

Nutritional status is an important predictor of wound healing potential (Hurd 2004; Leininger 2002; Medlin 2012). Malnutrition has been traditionally characterised by a person being underweight (NICE 2006), however research discusses the importance of considering the potential for malnutrition even in the obese person; for example Leininger 2002 highlights 120% of ideal body weight as a red flag warning for malnutrition. Malnutrition in an obese person may result in nutrient deficiencies due to an intake of high volume, poor quality food. Agte 2008 argues that in the general population, many people consume food that is either unhealthy or of poor nutritional value, containing high levels of fat, salt or sugar. These foods often lack the required proteins, vitamins, minerals and fibre, all of which are important in wound healing. Lonsdale 1994 proposes consideration of a theory called "high‐calorie malnutrition" in which he describes a state of excess intake of calories with concurrent nutrient deficiencies, resulting in an inadequate ability to utilise the calories efficiently. Kaidar‐Person 2008 also emphasises the importance of considering malnutrition in the obese person and highlights that research into the nutritional status of bariatric surgery patients suggests that many patients had a pre‐existing form of malnutrition prior to surgery. Indeed, Kaidar‐Person 2008 found that malnutrition in patients categorised as obese is common.

A high proportion of patients with DFUs are categorised as being obese. Indeed, Ribu 2007 found, when comparing a control group to both a non‐ulcer diabetic cohort and a DFU cohort, that the mean body mass index (BMI) of both diabetic cohorts was above 25 kg/m2. BMI is the ratio of weight to height, used in the prediction of mortality and morbidity (Eknoyan 2007). An individual with a BMI of above 25 kg/m2 is classed as being obese (WHO 1995).

UK guidelines on nutrition support in adults (NICE 2006) highlight poor wound healing as a clinical indicator of possible malnutrition, furthermore, the importance of nutrition in wound healing in general is well founded in the literature (Frias Soriano 2004; Lee 2006; Medlin 2012; Omote 2005). Leininger 2002 states that the main goal of nutrition in wound healing is to provide optimum calories and nutrition to aid healing, however deficiencies in protein, albumin, vitamin D, vitamin C and zinc have all been demonstrated to decrease wound healing rates.

Protein deficiency impairs wound healing by prolonging the inflammatory process, decreasing fibroblast (the collagen‐producing cell; they proliferate at the site of chronic inflammation (Marcovitch 2006)) and collagen development and reducing the supply of amino acids needed for healing (Demling 2009; Hurd 2004; Leininger 2002; Russell 2001). Heavily exuding wounds have an increased protein loss ranging from 2.1% to 13.8% (Liazka 2010) and thus the individual may require an increase in up to 3 grams per kilogram of protein intake per day for successful wound healing (Hurd 2004). Indeed, Pompeo 2007 found that enteral tube‐fed patients with wounds required 0.38 grams of protein per day more than patients without wounds to maintain nutritional status.

Albumin provides a source of amino acids necessary for a healing wound (Russell 2001). Furthermore, serum albumin is a traditional blood maker for nutritional deficiency (BDA 2001).

Vitamin A is important for the early inflammatory response and also aids in collagen and granulation tissue development (Demling 2009; Gottschlich 2001; Leininger 2002).

Vitamin C is important for wound healing as it is necessary for collagen synthesis (Demling 2009; Leininger 2002). As a result, vitamin C deficiency contributes to the development of fragile granulation tissue (Russell 2001). Ascorbic acid (or vitamin C) has shown a significant improvement in the healing rates of patients with pressure sores (Taylor 1974).

Zinc is required for cell division, protein synthesis and collagen deposition (Gottschlich 2001). Thus, a deficiency in zinc will reduce epithelial migration and impact negatively on tissue strength (Demling 2009). As zinc is transported by albumin, low albumin levels may impact on the measurable serum zinc levels (Desneves 2005). Furthermore, a deficiency in zinc can result in abnormal white cell function which increases the risk of wound infection (Demling 2009; Leininger 2002).

Iron enhances tissue regeneration because it is necessary for the transportation of oxygen to the tissues and for the formation of collagen (Gottschlich 2001; Leininger 2002; Russell 2001).

Nutrients are required for each phase of healing, for example, during the inflammatory phase low serum albumin (the major circulating protein (Hurd 2004)) will result in an inadequate inflammation resulting in impaired wound healing (Leininger 2002). Granulation tissue, which is formed during the proliferation stage, is largely composed of proteins of which collagen is in abundance. Indeed, collagen makes up 80% of the dry weight of the dermis and contributes to the wounds tensile strength (Martin 1992). Proteins and collagen are needed in the maturation stage to improve tissue strength (Medlin 2012).

Wounds require between 1.5 g and 3 g per kg per day of protein to ensure tissue regeneration (Hurd 2004; Medlin 2012) which is up to three times the normal protein intake (Hurd 2004). However, it should be noted that people with diabetes, and especially those with renal damage, should confine their intake of protein to reduce proteinuria and improve the prognosis regarding diabetic nephropathy (Zhang 2013). Indeed, the prevalence of renal impairment in people with diabetes is estimated at between 27.5% and 37.4% (deBour 2011; Ritz 1999; Rodriguez‐Poncelas 2013). Also of note, excessive dietary intake of vitamin A has been associated with foetal malformation (Azaïs‐Braesco 2000; Rothman 1995). Therefore, it remains important that people with DFUs receive adequate and correct nutrition in order to ensure successful closure of their DFU, whilst having regard for the potential complications arising from the presence of diabetes itself.

Why it is important to do this review

Altering nutritional intake has been shown in trials to improve wound healing in other wound types (Collins 2005; Haydock 1986) and seminal research has shown a link between nutritional status and DFU healing rates (Eneroth 2004). However, the precise role of nutrition in the treatment of DFU is as yet, unclear. The literature has shown that nutritional interventions may be beneficial in wound healing in general, for example, both EPUAP 2003 and Langer 2003 identify nutritional care as one of several interventions in the management of people with pressure ulcers. Nutritional intervention may potentially improve clinical outcomes such as healing rates and healing times of DFUs. However, as yet the role of nutritional interventions in people with DFUs has not been systematically reviewed. In order to provide clear guidance for the use of nutritional interventions in individuals with DFUs, it is important to systematically search and appraise the evidence. The outcomes of this review may provide evidence to formulate such guidance, furthermore, this review may indicate areas for future research. As DFUs have such a negative effect on the individual, the health service and society as a whole, treatments that enhance clinical outcomes in this patient group may also have an associated positive effect on health and social gain. It is for this reason that this review is necessary.

Objectives

To evaluate the effects of systemic nutritional interventions on the healing of foot ulcers in people with diabetes, in people of any age, in any care setting, with either type 1 or type 2 diabetes.

Methods

Criteria for considering studies for this review

Types of studies

Studies that randomise individuals (randomised controlled trials (RCTs)) or that randomise by groups (cluster‐RCTs) will be eligible for inclusion.

Types of participants

People of any age and sex, in any healthcare setting, with either type 1 or type 2 diabetes and an active foot ulcer.

Types of interventions

Intervention: nutritional supplementation (oral, enteral or parenteral nutrition) of any dose, or duration, or both, or special diet.

Comparison: comparisons between supplementary nutrition plus standard diet, versus standard diet alone, and between different types of supplementary nutrition (e.g. enteral versus parenteral).

Types of outcome measures

Primary outcomes

An objective measure of ulcer healing, such as:

  • time to complete healing;

  • absolute or percentage change in ulcer area or volume over time;

  • proportion of ulcers healed at the completion of the trial period; and

  • healing rate.

Secondary outcomes

An objective measure of:

  • costs of interventions;

  • quality of life as measured by a validated scale;

  • acceptability of the intervention (or satisfaction) with respect to patient comfort;

  • adverse events;

  • length of patient hospital stay;

  • development of any new foot ulcers;

  • amputation rate;

  • surgical interventions; and

  • osteomyelitis incidence.

Search methods for identification of studies

We will seek all RCTs and cluster‐RCTs which evaluate the use of systemic nutritional interventions in the treatment of foot ulcers in people with diabetes, using a search strategy designed by the Cochrane Wounds Group.

Electronic searches

We will search the following electronic databases.

  • The Cochrane Wounds Group Specialised Register.

  • The Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library latest issue).

  • Ovid MEDLINE (1946 to present).

  • Ovid MEDLINE (In‐Process & Other Non‐Indexed Citations).

  • Ovid EMBASE (1974 to present).

  • EBSCO CINAHL (1982 to present).

We will use the following provisional search strategy to search in The Cochrane Central Register of Controlled Trials (CENTRAL):.

#1 MeSH descriptor: [Foot Ulcer] explode all trees
#2 (diabet* near/3 ulcer*):ti,ab,kw
#3 (diabet* near/3 (foot or feet)):ti,ab,kw
#4 (diabet* near/3 wound*):ti,ab,kw
#5 (diabet* near/3 defect*):ti,ab,kw
#6 #1 or #2 or #3 or #4 or #5
#7 MeSH descriptor: [Nutrition Therapy] this term only
#8 MeSH descriptor: [Enteral Nutrition] this term only
#9 MeSH descriptor: [Parenteral Nutrition] explode all trees
#10 (nutrition* or diet* or enteral or parenteral or tube feeding):ti,ab,kw
#11 #7 or #8 or #9 or #10
#12 #6 and #11

We will combine the Ovid MEDLINE search with the Cochrane Highly Sensitive Search Strategy for identifying randomised trials in MEDLINE: sensitivity‐ and precision‐maximising version (2008 revision) (Lefebvre 2011). We will combine the EMBASE search with the Ovid EMBASE filter developed by the UK Cochrane Centre (Lefebvre 2011). We will combine the CINAHL searches with the trial filters developed by the Scottish Intercollegiate Guidelines Network (SIGN 2011). We will not impose any restrictions with respect to language, date of publication or study setting.

We will also search the following clinical trials registries.

Searching other resources

We will also search the bibliographies and reference lists of all retrieved and relevant publications identified by these strategies for further studies. We will contact manufacturers of nutritional interventions used in the treatment of wounds and experts in the field to ask for information relevant to this review.

Data collection and analysis

Selection of studies

Both review authors will independently assess titles and, where available, abstracts of the studies identified by the search strategy against the eligibility for inclusion in the review. We will obtain full versions of potentially relevant studies and both review authors will independently screen these against the inclusion criteria. Any differences in opinion will be resolved by discussion and, where necessary, reference to the Cochrane Wounds Group editorial base.

Data extraction and management

We will extract data from each article using a standardised data extraction sheet. Both review authors will independently extract data from eligible studies. Specifically, we will extract:

  • author;

  • title;

  • source;

  • date of study;

  • duration of study;

  • geographical location of study;

  • care setting:

  • inclusion/exclusion criteria;

  • sample size;

  • patient characteristics;

  • balance of groups at baseline;

  • study design details;

  • sources of funding;

  • study type;

  • method of randomisation;

  • allocation of concealment;

  • concurrent interventions;

  • wound status/category at baseline;

  • wound duration;

  • intervention details including type, dosage and duration;

  • control intervention details;

  • compliance;

  • outcome measures;

  • blinding (both patient and professional);

  • length of follow‐up;

  • loss to follow‐up;

  • results;

  • intention‐to‐treat analysis;

  • conclusions reported by study authors.

We will resolve disagreements by discussion or, where necessary, with reference to the Cochrane Wounds Group editorial base. We will enter and combine the data using Review Manager software (RevMan 2012).

Assessment of risk of bias in included studies

Both review authors will independently assess the included studies using The Cochrane Collaboration tool for assessing risk of bias (Higgins 2011a). This tool addresses six specific domains: sequence generation, allocation concealment, blinding, incomplete data, selective outcome reporting and other issues (Appendix 1). We will assess blinding and completeness of outcome data for each outcome separately. We will present our assessment of risk of bias using two 'Risk of bias' summary figures; one which is a summary of bias for each item across all studies, and a second which shows a cross‐tabulation of each trial by all of the risk of bias items. For trials using cluster randomisation, we will assess the risk of bias using the following domains: recruitment bias, baseline imbalance, loss of clusters, incorrect analysis and comparability with individually randomised trials (Higgins 2011b).

Measures of treatment effect

For dichotomous outcomes, we will calculate the risk ratio (RR) with 95% confidence intervals (CIs). For continuously distributed outcome data, we will use the mean difference (MD) with 95% CIs, if all trials use the same assessment scale. If trials use different assessment scales, we will use the standardised mean difference (SMD) with 95% CIs. We will report time‐to‐event data (e.g. time‐to‐complete wound healing) as hazard ratios (HRs) where possible, in accordance with the methods described in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2011).

Unit of analysis issues

Unit of analysis issues may arise from studies including participants with multiple DFUs, or in studies with individuals who are followed up and experience recurrence of DFUs. We will record whether trials presented outcomes in relation to a wound, a foot, a participant or as multiple wounds on the same participant. We will also record occasions where multiple wounds on a participant are (incorrectly) treated as independent within a study, rather than having within‐patient analysis methods applied. This will be recorded as part of the risk of bias assessment. For wound healing and amputation, unless otherwise stated, where the number of wounds appeared to equal the number of participants, we will treat the wound as the unit of analysis.

We will combine studies with multiple intervention groups into one group to create a simple pair‐wise comparison, however if there is no common effect between intervention groups, we may split the control group into two or more groups according to the number of intervention groups (Higgins 2011a).

Where possible, effect estimates and their standard errors from correct analyses of cluster‐randomised trials may be meta‐analysed using the generic inverse‐variance method in RevMan. Where a cluster‐randomised trial has been analysed on the individuals rather than the clusters we will approximate the correct analyses if possible using information on:

  • the number of clusters (or groups) randomised to each intervention group; or the average (mean) size of each cluster;

  • the outcome data ignoring the cluster design for the total number of individuals (e.g. number or proportion of individuals with events, or means and standard deviations);

  • and an estimate of the intracluster (or intraclass) correlation coefficient (ICC).

Dealing with missing data

It is common to have data missing from trial reports. Excluding participants post‐randomisation from the analysis, or ignoring those participants who are lost to follow‐up compromises the randomisation, and potentially introduces bias into the trial. In individual studies, where data on the proportion of ulcers healed are presented, we will assume that if randomised participants are not included in an analysis, their wound did not heal (i.e. they will be considered in the denominator but not the numerator). Where a trial does not specify participant group numbers prior to drop‐out, we will present only complete case data. In a time‐to‐healing analysis using survival analysis methods, drop‐outs should be accounted for as censored data. Hence all participants will contribute to the analysis. Such analysis assumes that drop‐outs are missing at random (i.e. not associated with time‐to‐healing). We will present data for area change, and for all secondary outcomes, as a complete case analysis.

Assessment of heterogeneity

We will assess clinical heterogeneity by examining potential influencing factors (e.g. care setting or wound stage). We will assess statistical heterogeneity using I2 statistic (Higgins 2003). This examines the percentage of total variation across studies due to heterogeneity rather than chance. Values of I2 over 75% indicate a high level of heterogeneity. We will carry out statistical pooling on groups of studies which are considered to be sufficiently similar. Where heterogeneity is absent or low (I2 = 0% to 25%) we will use a fixed‐effect model; if there is evidence of heterogeneity (I2 > 25%), we will use a random‐effects model. If heterogeneity is very high (I2 > 75%) we will not pool the data.

Assessment of reporting biases

Reporting biases arise when the dissemination of research findings is influenced by the nature and direction of results. Publication bias is one of a number of possible causes of 'small study effects', that is, a tendency for estimates of the intervention effect to be more beneficial in smaller RCTs. Funnel plots allow a visual assessment of whether small study effects may be present in a meta‐analysis. A funnel plot is a simple scatter plot of the intervention effect estimates from individual RCTs against some measure of each trial’s size or precision (Sterne 2011). We plan to present funnel plots for meta‐analyses comprising 10 RCTs or more using RevMan 5.3.

Data synthesis

Initially we will present a structured narrative summary of the studies reviewed. We will enter quantitative data into Review Manager software (RevMan 2012) and analyse it using the RevMan analysis software. If included studies are sufficiently similar in terms of population, inclusion criteria, interventions and outcomes (including the times at which outcomes are assessed in both intervention and treatment trials) we will consider pooling the data statistically, using meta‐analysis. We will use a fixed‐effect model if appropriate (i.e. I2 values are 25% or less), otherwise we will use a random‐effects model. We will not pool data where the I2 values are > 75%. We will include a summary of results from the data synthesis and assessment of quality of evidence in a 'Summary of findings' table for the main comparisons (Higgins 2011b). We will flag cluster trials in the review and text will be explicit regarding how data has been dealt with and presented. If it seems viable, we might combine data from cluster trials with individually randomised trials in the same meta‐analysis, but we would seek statistical advice to facilitate this.

Subgroup analysis and investigation of heterogeneity

If sufficient data are available we will undertake the following subgroup analyses:

  • Type of setting (community, hospital, inpatient, outpatient).

  • Type of intervention (oral, enteral, parenteral).

Sensitivity analysis

We will perform a sensitivity analysis by excluding studies of the lowest quality. In this sensitivity analysis, we will only include studies that are assessed as having a low risk of bias in all key domains, namely adequate generation of the randomisation sequence, adequate allocation concealment and blinding of outcome assessor, for the estimates of treatment effect.