Introduction

Since the time of Hippocrates, cachexia has been associated with a poor outcome in patients with cancer [1]. Indeed, cancer cachexia results in increased mortality rates, with up to 20% of cancer deaths related to malnutrition [2, 3].

Cachexia is not simply due to lack of adequate oral intake; rather, its pathophysiology is complex and includes a combination of systemic inflammation and hyper-metabolism [4]. This, in combination with decreased oral intake and reduced physical function, means that anabolism is impaired, resulting in loss of skeletal muscle.

With such a complex genesis, it may at first seem daunting to address these multiple components; however, there is a plausible argument that multimodal therapies targeting each of these elements, inflammation, decreased oral intake and reduced physical function, are necessary to optimally treat cachexia [5,6,7].

Appropriate nutritional intake is a key component of any intervention, and this has recently been emphasised by the European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines on cancer-related malnutrition and cachexia. It is now advocated that increased attention is paid to nutritional interventions for all patients with cancer [8].

Several key recommendations were subsequently made: nutritional intake should be screened regularly from the onset of cancer diagnosis, including those with advanced cancer; patients identified as having nutritional disturbance should undertake regular nutritional assessment including dietary intake, weight loss and body mass index [9].

In patients with cancer, the nutritional aim is often about maintaining or improving nutritional status, function and survival [10]. However, in patients with incurable cancer, the aim is often focused on improving quality of life and minimising symptoms such as nausea and vomiting which may impact on their nutritional intake [11].

However, the evidence to support regular nutritional assessment in patients with incurable cancer is not clear [9]. There is a need to collate and evaluate the evidence concerning the clinical consequences of nutrition support via dietary interventions including nutrition counselling with or without the use of oral nutritional interventions.

The aim of this systematic review was to assess the current evidence for nutrition support via nutritional interventions implemented in patients with incurable cancer.

Methods

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [12]. Ethical approval was not required.

Search strategy and selection criteria

Original studies with adult patients (> 18 years) with incurable cancer (defined as not curable but might receive antineoplastic treatment aimed at prolonging life and/or alleviate symptoms) and evaluating the effect of oral dietary interventions, were included. Eligible studies also had to have defined outcome measures such as body composition, including weight (measured in kg); pound or percent change in lean body mass (LBM); total body mass (TBM) or fat mass (FM); nutritional intake, including energy intake (measured as kcal, kJ or MJ, absolute intake and/or energy balance); and nutritional status, which were measured using validated tools such as the Patient Generated Subjective Global Assessment (PG-SGA) [13], and quality of life (QoL) was measured using patient-reported outcome measures such as the European Organisation for Research and Treatment of Cancer–Quality of Life-C30 (EORTC QLQ-C30) [14]. Both quantitative and qualitative trial designs were included.

Studies were excluded that evaluated the effect of either parental or enteral (via feeding tube) nutrition (including papers that evaluated mixed interventions that included enteral/parental nutrition). Studies were also excluded if the intervention selected nutritional compounds such as certain vitamins, fatty acids, proteins or amino acids. Case reports, conference abstracts, systematic reviews or studies with ten or less participants were not included. Language was limited to English only.

The literature search was conducted in the following electronic databases, MEDLINE, Embase and CINAHL, with all databases being searched from 1990 to 2018. The last search date was the 25th October 2018. The search was performed by an experienced librarian. The search strategy for all databases is reported in appendix 1 (supplementary material). Appropriate strategies were developed for each database.

Appraisal process

All titles retrieved from the literature search were reviewed (HB), and if potentially eligible, studies were retrieved in full and appraised independently (HB, BL and EH). If all three authors agreed that the studies met the eligibility criteria, these were then included in the review. Any disagreements regarding a trial were discussed between the three authors and a consensus was agreed upon. The PRISMA statement for reporting systematic reviews was used [12].

Eligible studies are summarised (Table 1) including risk of bias for each trial. Study quality was assessed by HB and CH using a modified risk of bias table, based on guidance from the Cochrane Handbook for Systematic Reviews of Interventions [15], and a summary table was developed (Table 2). The risk of bias for each patient-important outcome was evaluated and is presented in the table of modified summary of findings (Table 3).

Table 1 Study summaries
Full size table
Table 2 Quality of studies—risk of bias summary
Full size table
Table 3 Summary of findings (modified due to study types)
Full size table

Results

Search results and selection of studies

The literature search retrieved a total of 1139 papers (Fig. 1). After screening of the titles and identifying any duplicates, a total of 60 studies remained. One thousand and eighteen studies were removed at title. After reviewing each study against the eligibility criteria, 48 studies were excluded. Twelve studies were eligible, of which eight were randomised controlled trials (RCTs) [6, 16, 18, 20,21,22, 24, 26], three prospective observational studies [19, 23, 25] and one post hoc analysis study [17].

Fig. 1
figure 1

Literature search process

Full size image

Twelve studies assessing a total of 1266 patients investigated the effect of nutritional interventions in patients with incurable cancer. Predominant cancer types were gastrointestinal (including pancreatic and colorectal) and lung cancer, with over 40% receiving chemotherapy treatment.

Nutrition counselling with or without oral dietary intervention

Three studies (n = 438) examined nutrition counselling with or without oral dietary intervention, two RCTs [16, 22] and one prospective observational study [25]. The prospective observational study examined nutrition counselling alone, and the two RCTs examined nutrition counselling alongside an oral dietary intervention. One RCT compared the effects of nutrition counselling alone, the effect of oral nutritional supplements (ONS) alone, the effect of nutrition counselling and ONS in combination or the effect of no intervention [16]. The other RCT compared nutrition counselling and improved atta (IAtta) with nutrition counselling alone [22].

Findings by Kapoor et al. [22] reported that patients within the control arm had significantly decreased body weight (p = 0.003), mid-upper arm circumference (p = 0.002) and body fat (p = 0.002) by the end of the intervention. Although not significant, body weight gain was seen in the intervention group (p = 0.08); also, a significant increase in body fat (BF) (p = 0.002) was observed. Patients in the intervention group also reported a significant improvement in fatigue (p = 0.002) and appetite (p = 0.006).

Baldwin et al.’s [16] RCT was stopped early on advice of the independent data monitoring committee due to lack of efficacy. There was no significant difference in survival or QoL between the groups. Patients in the intervention group weighed more at one year than those in control group, but no difference was seen between those receiving ONS alone or the combination of ONS and dietary advice. There was no statistical difference between weight changes of non-survivors and survivors; however, less weight loss was seen in those who survived beyond 26 weeks.

Multimodal therapies alongside chemotherapy

Multimodal therapy, e.g. dietary intervention and physical exercise, delivered alongside chemotherapy was examined in five studies (n = 216) [6, 18, 23, 24, 26]; four were RCTs [6, 18, 24, 26] while the other was a prospective observational study [23].

Findings from Read et al. [23] saw improvements in body composition including a significant increase in mean weight at three weeks (p = 0.03) with this remaining stable up to week nine. Lean body mass (LBM) also maintained throughout the nine weeks. Significant improvements were also seen in energy levels (p = 0.03) between weeks three and nine, with all other QoL measures maintained. Dietary intake of n-3 fatty acids increased at week three and maintained up to week nine; this coincided with the commencement of the n-3 PUFA (polyunsaturated fatty acid)-enriched ONS. Significant improvement was seen in both eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) at week three and remained high up to week nine. Significant decreases in nutritional intake including protein (p = 0.003) and energy (p = 0.02) were seen following commencement of chemotherapy at the end of weeks three and nine.

Sanchez-Lara et al. [24] also saw improvements in body composition. Significant differences were noted between groups (p = 0.01) for LBM which increased in the intervention group but decreased in the control group. The intervention group also had significantly greater energy and protein intakes (p < 0.001) compared with those in the control group. No overall difference was seen in the response rate or survival between groups, but fatigue, neuropathy and loss of appetite decreased significantly in the intervention group (p = ≤ 0.05 for all).

Breitkreutz et al. [18] saw improvements in body composition including a fat free mass (FFM) increase in the intervention group compared with the control group. Body cell mass (BCM) decreased in the control group but was maintained within the intervention group, with significant intergroup differences between groups (p < 0.05). QoL was also seen to improve more in the intervention group compared with the control group, but the difference was not statistically significant.

Oral dietary interventions

Four studies examined the effect of oral dietary interventions alone (n = 611). Two were RCTs [20, 21], one exploratory prospective observational study [19] and one post hoc analysis study [17].

Fearon et al. [20] showed that although consumption was below the recommended dose, the intervention group still showed a significant correlation between ONS intake and improved body composition, including weight gain (p < 0.001) and an increase in LBM (p < 0.036). Weight gain was also associated with improved QoL in the intervention group (p < 0.01). However, no significant correlation was seen between intake and change in LBM between the two groups (p < 0.043). Increased plasma EPA levels were also associated with weight and LBM gain (p < 0.01).

Casas et al. [19] showed significant differences in the intervention group for anxiety (p = 0.023) and depression (p = 0.011), with QoL showing significant differences from baseline measurements between groups (p = 0.017). Significant differences were also seen between the groups in the global scale (p = 0.016) and fatigue scale (p = 0.007).

Summary of findings

Twelve studies were identified, all from the outpatient setting. Following assessment of study quality using a modified risk of bias table, based on guidance from the Cochrane Handbook for Systematic Reviews of Interventions [15], we have shown that there is moderate quality evidence to support the need for increased attention to nutrition support in patients with incurable cancer; however, despite some statistically significant results being reported, the clinical effects of them were small.

Body composition

Eleven studies examined body composition as an outcome measure. Six reported an increase in weight [6, 17, 18, 22, 23, 26] of which one, looking at a combination of nutrition counselling and ONS alongside chemotherapy, reported an increase in FFM (p < 0.05) and maintenance in BCM compared with the control group [18]. One study, examining the effects of IAtta, reported an increase in body weight gain (p = 0.08) with a significant increase in body fat (p = 0.002) [22]. Only one study, examining nutrition counselling with the emphasis on restricting carbohydrates, reported significant weight loss following intervention [25]. Four studies reported weight stability [16, 20, 21, 24], although one of those studies, examining nutrition counselling and consumption of an n-3 PUFA-enriched ONS alongside chemotherapy, reported an increase in LBM [24]. Of the six studies reporting an improvement in body composition, one study examined nutrition counselling alongside dietary intervention of IAtta [22], one study examined the effect of an n-3 PUFA-enriched ONS [17] and four studies examined multimodal therapies alongside chemotherapy [6, 18, 23, 26]. Of the six studies which saw improvements, all examined an oral dietary intervention, of which five examined ONS [6, 17, 18, 23, 26], with three of those being an n-3 PUFA-enriched ONS, and one examined IAtta [22].

QoL

Eleven studies examined QoL as an outcome measure with eight studies reporting an improvement in QoL, measured on various subscales (three studies saw significant improvements [19, 22, 24], and five studies saw non-significant improvements) [17, 18, 20, 25, 26], compared with the control group and three studies reporting no difference between groups [16, 21, 23]. Of the eight studies reporting an improvement in QoL, two were examining the effect of nutrition counselling alongside a dietary intervention [22, 25], three examined an oral nutritional intervention [17, 19, 20], including one examining ice cream as a dietary intervention compared with ONS [19]. Three studies examined multimodal therapies alongside chemotherapy [18, 24, 26]. Of the eight studies which saw improvements in QoL, seven examined oral dietary interventions, of which six examined ONS [17,18,19,20, 24, 26], with four of those being an n-3 PUFA-enriched ONS, and one study examined IAtta as a dietary intervention [22]. The remaining study was examining nutrition counselling aimed at restricting carbohydrates [25].

Nutritional intake

Nine studies examined nutritional intake as an outcome measure, with six studies reporting an improvement in nutritional intake [17, 18, 20,21,22, 24] including protein and energy intake (p < 0.01) and three studies reporting a reduction in appetite loss [21, 22, 24]. Only one of these studies, examining nutrition counselling and consumption of an n-3 PUFA-enriched ONS alongside chemotherapy, reported a decrease in intake following commencement of chemotherapy [23]. One study, examining nutrition counselling alongside an oral dietary intervention, failed to analyse nutritional intake due to compliance issues with the outcome tool used [16], and one study, examining nutrition counselling and consumption of an n-3 PUFA-enriched ONS alongside chemotherapy, showed no difference between groups [26]. Of the six studies reporting an improvement in nutritional status, one examined nutrition counselling alongside dietary intervention [22], three examined an oral dietary intervention [17, 20, 21] and two examined multimodal therapies alongside chemotherapy [18, 24]. Of the six studies which saw improvements, all examined an oral dietary intervention: four were examining ONS [17, 18, 20, 24], including three of those examining n-3 PUFA-enriched ONS, one study examined IAtta as a dietary intervention [22] and the final study examined ice cream as a dietary intervention [19]. Of the six studies which saw improvements in nutritional intake, three of these also saw improved QoL [18, 22, 24], with three studies seeing improvements in body composition including weight [17, 18, 22], free fat mass [18] and body fat [22].

Nutritional status

Six studies examined nutritional status as an outcome measure with three studies seeing improvements in nutritional status [19, 22, 24]. Three studies reported no differences between groups [6, 23, 26]. Of the three studies that reported improvements in nutritional status, one study examined ice cream as a dietary intervention compared with ONS [19], one study examined nutrition counselling alongside the addition of IAtta [22] and one study examined nutrition counselling and the consumption of an n-3 PUFA-enriched ONS alongside chemotherapy [24]. Of the three studies which saw improvements, all examined an oral dietary intervention: one examined an n-3 PUFA-enriched ONS [24], one examined IAtta [22] and the last one examined ice cream as a dietary intervention, compared with ONS [19].

Discussion

This systematic review has found only limited evidence as to the most effective nutrition intervention for patients with incurable cancer, despite various guidelines [8, 27].

The National Institute for Clinical Excellence (NICE) [27] guidelines and the ESPEN guidelines [8] highlight the need for early nutritional screening in order to identify patients who are malnourished. Diagnostic criteria for cachexia have been developed and used to classify patient’s degree of cachexia; these consider food intake, catabolic drivers, muscle mass/strength and effect of cachexia on the patient [4]. Cachexia classifications highlight that if cachexia is present, it can develop progressively from pre-cachexia to cachexia and to refractory cachexia which cannot be fully reversed by conventional nutrition support and leads to progressive functional impairment. The studies included in this review did not classify the stage of cancer cachexia in which interventions were delivered. It would be interesting in future work to assess optimal timings of delivery of nutritional interventions. Should the cachexia classification criteria be used routinely for cancer patients, alongside nutritional assessments, in order to identify as early as possible those who are not only malnourished but also at risk of cachexia and to what degree? This should be considered for patients when they are initially diagnosed with cancer and regularly screened throughout their cancer journey to minimise the risk of developing malnutrition/cachexia complications and/or prevent further deterioration which may impact on their functional status [28].

Nutrition counselling is considered the most appropriate first-line nutritional intervention [8, 27], and the findings herein support this. Further aspects however need to be considered including who is the best person to conduct the nutritional intervention, when should this take place, and should advice be standardised [29]. Patients are often provided with nutritional advice at varying time points of their journey from different health professionals, and advice can often be conflicting or incorrect. Symptoms as a result of deteriorating status or from cancer treatment also need to be taken into consideration when providing nutritional advice as these can often have a negative effect on oral intake [30]. The type of interventions within studies should therefore be clearly described for both the control and the intervention groups as well as timeframes undertaken.

High attrition rates are common in studies involving palliative care patients, and this was also evident in the studies examined in this review, with attrition rates over 40% recorded in three studies [17, 20, 22]; this is often due to the frail nature of this patient group leading to withdrawal and high dropout rates [31].

Appropriate outcome measures also need to be considered, and it is imperative in palliative care that these are relevant to assess appropriate palliative goals of care. Various tools have been developed such as the PG-SGA [13] to measure nutritional outcomes, but there is no defined consensus on which tools are most appropriate. Due to the nature of this patient group, patients can often be too unwell, frail or fatigued to complete self-completed measurements. This can lead to reporting bias whereby frailer patient data is not included or missing [32, 33].

Limitations

Relevant studies may have been missed in this review, despite a thorough search strategy being implemented; however, we believe we identified all appropriate studies. Meta-analysis of studies was not possible due to the differences in trial designs. Multiple assessors assessed study quality to limit any risk of bias, and any discrepancies were discussed in detail and then a consensus was agreed upon. Baldwin et al. [16] highlights that although RCTs are the gold standard, these are difficult to undertake for nutritional intervention studies. They argue it is often impossible to blind both the participants and the person undertaking the intervention or to have a placebo for the control group which can often then lead to bias [29]; indeed, this was the case for most of the studies included in this review. This study also reviewed observational studies, which are often seen as inferior to RCTs due to high risk of confounding factors and selection bias of patients [34].

Conclusion

This review demonstrates moderate evidence for nutrition support in patients with incurable cancer, which supports the recommendations by ESPEN for increased attention to nutritional support in this patient group. Further high-quality studies are needed in order to identify the most appropriate types of nutritional interventions.