Dietary and nutrition interventions for the therapeutic treatment of chronic fatigue syndrome/myalgic encephalomyelitis: a systematic review.

BACKGROUND: Chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) is characterised by unexplained fatigue for at least 6 months accompanied by a diverse but consistent set of symptoms. Diet modification and nutritional supplements could be used to improve patient outcomes, such fatigue and quality of life. We reviewed and discussed the evidence for nutritional interventions that may assist in alleviating symptoms of CFS/ME.
METHODS: Medline, Cinahl and Scopus were systematically searched from 1994 to May 2016. All studies on nutrition intervention were included where CFS/ME patients modified their diet or supplemented their habitual diet on patient-centred outcomes (fatigue, quality of life, physical activity and/or psychological wellbeing).
RESULTS: Seventeen studies were included that meet the inclusion criteria. Of these, 14 different interventions were investigated on study outcomes. Many studies did not show therapeutic benefit on CFS/ME. Improvements in fatigue were observed for nicotinamide adenine dinucleotide hydride (NADH), probiotics, high cocoa polyphenol rich chocolate, and a combination of NADH and coenzyme Q10.
CONCLUSIONS: This review identified insufficient evidence for the use of nutritional supplements and elimination or modified diets to relieve CFS/ME symptoms. Studies were limited by the number of studies investigating the interventions, small sample sizes, study duration, variety of instruments used, and studies not reporting dietary intake method. Further research is warranted in homogeneous CFS/ME populations.

Introduction

Chronic fatigue syndrome, also known as myalgic encephalomyelitis (CFS/ME), is a chronic, disabling illness characterised by unexplained persistent and debilitating fatigue, and this is accompanied by a diverse but consistent set of symptoms. CFS/ME has an unknown aetiology, as well as no known specific pathogenesis; therefore, there is no diagnostic pathology test. Rather, diagnosis is one of disease exclusion, and is made in accordance with symptom‐specific criteria 1, 2. Prior to year 1994, many case definitions were used in research to aid in the diagnosis of this condition, thus limiting comparisons of published studies 3. The Fukuda (1994) criteria was established to overcome inconsistency in the application of case definitions, as well as to assist in defining a distinct group of cases 2, and is now the most frequently used case definition. To meet the Fukuda 2 criteria, a patient must have debilitating unexplained fatigue present for at least 6 months that is not explained by ongoing exertion or medical or psychiatric conditions, and is not alleviated by rest. In addition, the fatigue must be accompanied by four or more of the following symptoms: post‐exertional malaise, difficulty with short‐term memory or concentration, unrefreshed sleep, sore throat, muscle and/or joint pain, headaches, and tender lymph nodes 2. A more recent, and alternative diagnostic criteria for CFS/ME, is the International Consensus Criteria (ICC) 1, which identifies distinct CFS/ME symptoms associated with neurological, immunological, gastrointestinal and energy production impairments 1.

The current worldwide prevalence of CFS/ME is estimated to be between 0.8% and 3.3% 4. In the USA in 2008, the treatment and management of CFS/ME was estimated to cost US$319 million annually, with a direct cost of US$7406 per patient 5. This syndrome is heterogeneous in nature, with CFS/ME patients reporting different accompanying symptoms, at different severities 1, frequency (continuous or intermittent symptoms) and duration 6. Many CFS/ME patients experience significant cognitive and physical impairment and, consequently, a substantial decline in social, occupational, educational and personal activity 7. Thus, CFS/ME significantly affects and interferes with everyday life and patients relationships.

Many CFS/ME patients complain of gastrointestinal symptoms, including but not limited to early satiety, abdominal distension and/or pain, nausea, vomiting and altered bowel habits 8, 9. Additionally, irritable bowel syndrome (IBS) 8, 9, 10, 11, a functional disorder of the gastrointestinal tract, coeliac disease and food intolerance (e.g. wheat and dairy) 9 are frequently observed in CFS/ME patients.

CFS/ME patients report a high use of nutritional supplements 12, 13 and approximately 50% of patients have reported food intolerances 9, 10, 11, 12 and benefit from dietary modification 12. Therefore, dietary therapy, including diet modification and/or provision of dietary supplements, may be beneficial in alleviating symptoms and reducing fatigue in CFS/ME. The present study aimed to systematically review original research investigating nutrition interventions in the symptom management of CFS/ME patients measured using patient‐centred outcomes including fatigue, quality of life, physical activity and/or psychological wellbeing).

Materials and methods

Literature search

Three databases were utilised: Medline (EBSCOhost), Cinahl (EBSCO) and Scopus. The following terms were systematically searched as full‐text and Medical Subject Headings (MeSH) terms (Medline and Cinahl): syndrome, chronic fatigue (which includes chronic fatigue syndrome and myalgic encephalomyelitis) and food, diet, nutrition therapy, diet therapy, vitamins, minerals, micronutrients, dietary supplements and/or nutritional supplements. Search results were limited to English language, publication date (year 1994–2016) and humans (all databases except Scopus). A secondary search was completed, whereby included studies and review articles were reviewed for forward citations and to identify other eligible studies. The final search was completed on 6 May 2016.

Inclusion and exclusion criteria

Studies that fulfilled the following criteria were eligible for inclusion: (i) all studies that were intervention research, defined as studies that evaluated the effectiveness of food and/or nutritional supplement on outcome measures; (ii) CFS/ME diagnosis according to Fukuda 2, Canadian (2003) 14 or International Consensus Criteria (ICC) (2011) 1; (iii) adults aged 18 years and over; (iv) studies that had accessible full‐text articles written in English; (v) year searched 1994 to present to exclude earlier studies prior to 1994 Fukuda criteria 2, 15; and (vi) studies comprising journal articles based on original research. The primary outcome of interest for this review was fatigue. Secondary outcomes evaluated were quality of life, physical activity and psychological wellbeing. Studies were excluded if they explicitly combined CFS/ME with other patient groups [e.g. CFS/ME and fibromyalgia (FMS)]. Although CFS/ME often co‐occurs with FMS and other disorders such as IBS 1, the co‐occurrence of FMS was excluded to understand the effect of nutrition interventions on outcome measures specifically in CFS/ME. Studies that used multi‐treatments (e.g. nutrition and pharmaceutical treatment), duplicate studies, case reports/studies or review articles and studies not meeting the above inclusion criteria were also excluded.

Selection of studies

Titles and abstracts for each article were initially screened on the basis of eligibility criteria. Two review authors independently assessed full‐text articles for suitability for inclusion in this review, and study quality, followed by a research meeting of all team members that confirm articles for inclusion in this review.

Data extraction and quality assessment

Eligible studies were read and the relevant data were extracted (Tables 1, 2, 3) including: (i) study design; (ii) CFS/ME case definition; (iii) country; (iv) sample size; (v) age of participants (vi) sex, percentage of female participants; (vii) illness duration; (viii) body mass index; (ix) weight; intervention duration; washout period between trials; (x) nutrition intervention(s) being investigated; (xi) name of instrument to evaluate study outcomes; and (xii) result of intervention and level statistical significance.

Table 1

Summary of participant and study characteristics of the included studies

Reference	Study design	Dx	Country	Sample (n)		Age (years), mean (SD)		Sex, female %		Illness duration, Mean (SD)		BMI (kg m−2), mean (SD)		Weight (kg), mean (SD)
				Tx	Con	Tx	Con	Tx	Con	Tx	Con	Tx	Con	Tx	Con
Fukuda et al. 15	RCT, PAR	F	Japan	17	14	34.8 (9.36)	39.5 (8.50)	76	86	NR	NR	NR	NR	NR	NR
Fukuda et al. 15 (Pilot)	Cohort, OPT	F	Japan	20	–	36.8 (6.88)	–	75	–	123 (64.2)	–	NR	–	NR	–
Ostojic et al. 26	RCT, CO	F	Serbia	21	–	39.3 (8.8)	–	100	–	–	–	NR	NR	62.8  (8.5)	–
Castro‐Marrero et al. 21	RCT, PAR	F	Spain	39	34	NR	NR	100	100	15.4 (8.9)	14.7 (6.2)	26.7 (5.2)	25.9 (2.4)	68.5 (14.6)	72.1 (13.7)
Witham et al. 24	RCT, PAR	F, C	Scotland	25	25	48.1 (12.0)	50.7 (13.1)	72	80	NR	NR	28.8 (7.9)	29.8 (5.4)	NR	NR
Maric et al. 30	Cohort, PRO	F	Serbia	38	–	NR	–	100	–	NR	–	NR	–	NR	–
Sathyapalan et al. 32	RCP, CO	F	UK	10	–	52 (8)	–	60	–	NR	–	28.3 (2)	–	NR	–
Sullivan et al. 33	Cohort, OPT	F	Sweden	15	–	43	–	67	–	NR	–	NR	–	NR	–
Rao et al. 31	RCP, PAR	Ca	Canada	19	16	NR	NR	NR	NR	NR	NR	NR	NR	NR	NR
Hobday et al. 27	CS, NC	F	UK	25	27	44 (10.2)	42.3  (11.9)	88	78	NR	NR	NR	NR	NR	NR
The et al. 23	RCT, PAR	F	NLD	30	27	40.9 (9.4)	43.4  (11.2)	77	59	NR	NR	NR	NR	NR	NR
McDermott et al. 22	RCT, PAR	F	UK	37	34	42 (15)	43 (12)	76	68	NR	NR	NR	NR	NR	NR
Vermeulen et al. 28	CS, NC	F	AMS	30	30 30	37 (11)	38 (11) 42 (12)	77	77 77	5.5 (1–23)a	3.0 (0.5–25)a 6.0 (1.0–21)a	NR	NR	NR	NR
Brouwers et al. 20	RCT, PAR	F	NLD	27	26	40.0 (9.9)	38.9 (10.9)	74	65	8.0 (2–15)a	4.5 (2–10)a	NR	NR	NR	NR
Ockerman et al. 25	RCT, CO	F	Sweden	22	–	50	–	86	–	NR	–	NR	–	NR	–
Plioplys et al. 29	CS, NC, CO	F	USA	28	–	40 (18–67)a	–	57	–	5.0 (1–20)a	–	NR	–	NR	–
Forsyth et al. 6	RCT, CO	F	USA	26	–	39.6	–	65	–	7.2	–	NR	–	NR	–

AMS, Amsterdam; BMI, body mass index; C, Canadian case definition 14; Ca, Carruthers clinical case definition; CO, cross‐over design; Con, control group; CS, comparative study design; Dx, case definition for CFS/ME diagnosis; F, Fukuda (1994)2; NC, no comparative controls; NLD, Netherlands, NR, not reported; OPT, Open labelled pilot trial; PAR, parallel design; PRO, prospective design; RCP, randomised control pilot study; RCT, randomised control study; Tx, treatment group; –, not applicable.

Median (range) in years.

Table 2

Study intervention characteristic and results for the primary outcome of fatigue

Reference	Duration (weeks)a	Washout periodd	Treatment intervention	Control intervention	Fatigue outcome	Fatigue result
Fukuda et al. 15	12	–	Ubiquinol‐10 (150 mg day−1 post meal)	Placebo	Chalder Fatigue Scale	NS
Fukuda et al. 15 (Pilot)	8	–	Ubiquinol‐10 (150 mg day−1 post meal)	–	Chalder Fatigue Scale	NS
Ostojic et al. 26	3 months	2 months	GAA (2.4 g day–1)	Placebo (cellulose)	Multidimensional Fatigue Inventory	NS for general or physical fatigue subscales. Significantly different for mental fatigue, activity and motivation (P < 0.05).
Castro‐Marrero et al. 21	8	–	NADH (200 mg day−1) + CoQ10 (20 mg day−1)	Placebo	Fatigue Index Symptom Questionnaire (FIS‐40)	Decreased (P < 0.05)
Witham et al. 24	6 months	–	Vitamin D3 (100 000 units orally every 2 months)	Placebo	Piper Fatigue Scale	NS
Maric et al. 30	2 months	–	Multivitamin mineral supplement (Supradyn®)	–	Fibro Fatigue Scale (FFS)	Unclear, inconsistent data reported for total FFS. Fatigue subscale significant (P = 0.0009)
Sathyapalan et al. 32	8	2 weeks	High cocoa liquor/ polyphenols rich chocolate (85% cocoa solids; 2.28 MJ/100 g; 15 g bar three times daily)	Cocoa liquor free/low polyphenols chocolate (2.29 MJ/100 g; 15 g bar three times daily)	Chalder Fatigue Scale	Decreased (P = 0.01)
Sullivan et al. 33	4	–	Probiotic bacteria (2 dL of Cultura Dofilus natural yogurt twice daily for 30 days. Lactobacillus F19, Lactobacillus acidophilus NCFB 1748 and Bifidobacterium lactis Bb12)	–	Visual analogue Scales for general, muscle and neurocognitive fatigue	NS for general fatigue. Improved neurocognitive function (P = 0.040)
Rao et al. 31	8		Probiotic bacteria (8 billion cfu of Lactobacillus casei, Shirota)	Placebo	NR	NR
Hobday et al. 27	24	–	Low sugar low yeast diet + consult at baseline and 24 weeks + phone call monthly	Healthy eating dietb + consult at baseline and 24 weeks + phone call monthly	Chalder Fatigue Scale	NS
The et al. 23	14		Acclydine	Placebo	CIS‐fatigue Daily Observed Fatigue	NS
McDermott et al. 22	8	–	BioBran MGN‐3 (6 g; 2 g three times per day dissolved in water or milk)	Placebo	Chalder Fatigue Scale	NS
Vermeulen et al. 28	24	–	Acetyl‐l‐carnitine (ALC) (2 g day1; post BF)	Propionyl‐l‐carnitine (PLC) (2 g day1; post BF) OR ALC + PLC (2 g ALC + 2 g PLC day1; post BF)	Multidimensional Fatigue Inventory	Within group analysis showed that ALC significantly improved mental fatigue (P = 0.015) PLC and ALC + PLC significantly improved general fatigue.
Brouwers et al. 20	10	–	Multinutritional supplement, twice daily	Placebo	CIS‐fatigue Daily Observed Fatigue	NS
Ockerman et al. 25	3 months	2 weeks	Pollen and pistil extract (Polbax) (7 ×  tablets in 1 dose at BF)	Placebo	Subjective Symptom	No between group analyses. Decrease in self‐reported symptoms in treatment group
Plioplys et al. 29	8	2 weeks	L‐carnitine (1 g three times daily)	Amantadinec (100 mg once daily (morning) for 4 weeks and then increased to 100 mg twice daily (morning and afternoon)	Fatigue Severity Scale	NS
Forsyth et al. 6	4	4 weeks	NADH (10 mg day−1 with water, 45 min before BF, fasted)	Placebo	50 item questionnaire based on CDC criteria for CFS/ME	No between group comparisons. Within group analysis showed that 31% & 8% of patients benefited from NADH versus PBO, respectively (P < 0.05)

BF, Breakfast; CDC, Centers for Disease Control and Prevention; CFS/ME, chronic fatigue syndrome/myalgic encephalomyelitis; CFU, colony‐forming units; CIS‐Fatigue, Checklist Individual Strength subscale fatigue severity; CoQ10, coenzyme Q10; GAA, guanidinoacetic acid; NADH, nicotinamide adenine dinucleotide hydride; NR, not reported; NS, not significant; PBO, placebo; VAS, Visual Analog Scale.

Duration reported for cross‐over trials is duration of each intervention, excluding washout period.

Healthy eating diet as per Department of Health guidelines (COMA, 1991) for the general population.

Results are outside the scope of this systematic review.

Washout period between crossover groups.

Table 3

Secondary outcome measure of psychological wellbeing, quality of life and physical activity level

Reference	Secondary outcome measure(s)	Results
Psychological wellbeing
Fukuda et al. 15	CES‐D	NS
Fukuda et al. 15 (Pilot)	CES‐D	NS Improvement in depressive symptoms was dependant on increase in total plasma CoQ10 levels
Witham et al. 24	HADS	NS
Sathyapalan et al. 32	HADS	Significant decrease in anxiety and depression (P = 0.01)
Rao et al. 31	Beck Depression Inventory Beck Anxiety Inventory	NS Significant decrease in anxiety symptoms (P = 0.01)
Hobday et al. 27	HADS	NS
McDermott et al. 22	HADS	NS
Plioplys et al. 29	Beck Depression Inventory Symptom Checklist 90‐R	Significant decrease in depression at 8 weeks (P = 0.22) Significant decrease in somatisation (P = 0.012), obsessive‐compulsive (P = 0.036), anxiety (P = 0.006) and depression (P = 0.006) subscale; and all summary scales including GSI (P = 0.007), PSDI (P = 0.000) and PSTI (P = 0.038) at 8 weeks
Quality of life
Ostojic et al. 26	SF‐36	Significant improvement in physical (P = 0.04) and mental common scores (P = 0.00)
Maric et al. 30	SF‐36	NS
Sathyapalan et al. 32	London Handicap Scale	Significant increase in residual function (P = 0.01)
Sullivan et al. 33	SF‐12 Health Survey	NS
Hobday et al. 27	MOS SF‐36	NS
The et al. 23	SIP‐8 score	NS
McDermott et al. 22	WHO QOL‐BREF	Social wellbeing subscale only (P = 0.02)
Vermeulen et al. 28	CGI	Within group analysis showed improvement in ALC and PLC but not ALC + PLC
Brouwers et al. 20	CDC symptom checklist SIP‐8 score	NS NS
Plioplys et al. 29	CFS Impairment Index CFS Severity Index	Within group analysis showed significant improvement in total function (P = 0.001) and physical (P = 0.000) and mental (P = 0.038) subsets at 8 weeks Significant improvement in function (P = 0.031)
Physical activity level
Ostojic et al. 26	Actigraphic assessment Energy expenditure, duration and intensity	NS NS
The et al. 23	Actigraphic assessment	NS
Brouwers et al. 20	Actigraphic assessment	NS

CDC, Centers for Disease Control and Prevention; CES‐D, Centre for Epidemiologic Studies Depression Scale; CGI, Clinical Global Impression of Change; CoQ10, coenzyme Q10; GSI, Global Severity Index; HADS, Hospital Anxiety and Depression Score; LHS, London Handicap Scale; MOS SF‐36, Medical Outcomes Survey Short Form; NS, no significant difference; PA, physical activity; PSDI, Positive Symptom Distress Index; PSTI, Positive Symptom Total Index; SF‐36, Quality of Life Scale; SIP‐8, Sickness Impact Scale.

To evaluate study quality and bias, the Rosendal scale 16, which combines the PEDro scale 17, Jadad scoring system 18 and Delphi List 19, was utilised (see Supporting information, Table S1). Items 15 and 16 of the Rosendal scale were excluded because outcomes associated with exercise performance (e.g. VO2 max) were not relevant to this review. Item 15 was replaced to included assessment of if an appropriate washout period was used for cross‐over trials because of its relevance to nutrition interventions. A Rosendal score cut‐off of 60% is classified as excellent methodological quality 16.

Results

Overview of studies and study quality

Figure 1 shows the PRISMA flow diagram with the number of included and excluded studies. Seventeen studies were included in this systematic review of nutrition and dietary interventions in CFS/ME patients. The included studies were six randomised control trial (RCT) parallel designs 15, 20, 21, 22, 23, 24; three RCT cross‐over designs where patients were their own control 6, 25, 26; three comparative studies without concurrent control (two‐ or three‐arm parallel groups) 27, 28, 29; one prospective cohort study (before and after) 30; and four pilot studies 15, 31, 32, 33. All RCT (parallel and cross‐over) 6, 15, 20, 21, 22, 23, 24, 25, 26 and two pilot studies 31, 32 reported using a double‐blind design. Studies varied in study quality, with the Rosendal score ranging from 10% to 86%. Eleven of the 17 studies (65% of included studies) had a Rosendal score of >60% 6, 15, 20, 21, 22, 23, 24, 26, 27, 28, 32 (see Supporting information, Table S1).

Figure 1

Flow diagram of database and secondary search for included studies in the review of nutrition intervention on chronic fatigue syndrome/myalgic encephalomyelitis (CFS/ME) outcomes.

Participant and study characteristics

The participant characteristics of the included studies are summarised in Table 1. For the diagnosis of CFS/ME, 15 studies used the Fukuda (1994) case definition 6, 15, 20, 21, 22, 23, 25, 26, 27, 28, 29, 30, 32, 33, one study used Carruthers (2003) 31 and one combined Fukuda (1994) and the Canadian Clinical Case Definition 24. Two studies also included participants with high fatigue severity measured with subjective fatigue subscale of the Checklist Individual Strength (CIS‐fatigue ≥ 40) 20 and Chandler Fatigue Scale (score 10/11) 32. The mean sample size for each study was approximately 24 participants.

All studies, except two, reported fatigue as the primary outcome, and one study did not report fatigue at all 31. Of the 17 studies, five studies used the Chalder Fatigue Scale 15, 22, 27, 32, three studies used an investigator designed symptom questionnaire 6, 25, 33, two studies used the Multidimensional Fatigue Inventory 26, 28, and two studies used both the Checklist Individual Strength Fatigue Subscale and Daily Observed Fatigue 20, 23. All other studies used either the Fatigue Index Symptom Questionnaire (FIS‐40) 21, Fatigue Severity Scale 29, Fibro Fatigue Scale 30 and Piper Fatigue Scale 24. No trials reported a dietary intake method (diet record or food frequency questionnaires) for measurement at baseline or at the end of the study period to determine whether dietary changes occurred.

Interventions on fatigue

In total, 14 different interventions were evaluated (Table 2). All randomised control trials, including pilot studies, compared a form of nutritional supplement with a placebo control 6, 15, 20, 21, 22, 23, 24, 25, 26, 31. One RCT compared the combined effect of nicotinamide adenine dinucleotide hydride (NADH) and coenzyme Q10 (CoQ10) with a placebo, and found a significant reduction of fatigue after 8 weeks of treatment compared to placebo (P < 0.05) 21. Another RCT observed a significant decrease in mental fatigue after guanidinoacetic acid (GAA) supplementation compared to a placebo 26. One RCT tested the effect of pollen and pistil extract and a placebo after 3 months of supplementation on fatigue 25. In the treatment group, a significant decrease in self‐reported symptoms was observed, whereas no significant difference was observed for the placebo group from baseline 25. A randomised control pilot study reported a significant decrease in self‐reported fatigue after participants consumed high cocoa liquor (polyphenol rich chocolate) compared to iso‐caloric chocolate for 8 weeks (P = 0.01) 32. Other RCT either observed no significant difference between the nutrition intervention and placebo control 15, 20, 22, 23, 24 or did not report between group analysis 6, 25. One cohort study observed significant improvement in neurocognitive fatigue (P = 0.040) but not in general fatigue after probiotic bacteria supplementation 33. All other studies reported no difference either compared to a comparative intervention 27 or after the intervention 15, 29, 30.

Interventions on secondary outcomes

Of the 17 studies, 14 studies evaluated the effect of the nutrition intervention on either quality of life (n = 10) 20, 22, 23, 26, 27, 28, 29, 30, 32, 33, physical activity (n = 3) 20, 23, 26 and/or psychological wellbeing (n = 8) 15, 22, 24, 27, 29, 31, 32 (Table 3). Three studies observed improvement in psychological wellbeing 29, 31, 32. One RCT study observed decrease in anxiety but not depression following probiotic supplementation 31. Two comparative studies reported significant improvement in both anxiety and depression after consumption of polyphenol rich chocolate 32 and l‐carnitine 29 compared to baseline. Consumption of l‐carnitine also improved five other psychometric tests 29. Two RCT observed significant difference in quality of life for supplementation of BioBranTM MGN‐3 (Daiwa Pharmaceutical, Tokyo, Japan) (social wellbeing subscale only) 22 and guanidinoacetic acid 26. Two comparative studies reported improvement in quality of life after supplementation of acetyl‐l‐carnitine and propionyl‐l‐carnitine 28, as well as l‐carnitine 29. No studies reported improvement in physical activity level; however, one study reported significant improvement in quadriceps isometric strength and VO2 max after 3 months of supplementation of guanidinoacetic acid in patients compared with controls 26. All other studies observed no difference of the nutrition intervention on any of the secondary outcomes 15, 20, 23, 27, 30, 33.

Discussion

Elimination diets, dietary restriction and the addition of nutritional supplements to habitual diet are widely reported to be used by CFS/ME patients, and many individuals claim beneficial effects of these dietary interventions in reducing fatigue 12. This systematic review has summarised the available evidence on dietary and nutritional interventions in CFS/ME on patient outcomes, including fatigue, quality of life, psychological wellbeing and physical activity level. Additionally, this review has updated the literature search according to scientific accepted diagnostic criteria 34. Our study highlights methodological limitations in the evidence and an overall lack of evidence that explores the therapeutic effect of diet and nutritional supplementation in CFS/ME.

Participants in this review were predominantly female, with a mean age of 35–50 years, residing in Europe. This finding is relatively consistent with epidemiological studies that report a higher prevalence of CFS/ME in females 7, 9, 35 and individuals aged 35 and 45 years 7, 9. However, this sex imbalance may be a consequence of studies recruiting participants from clinics or universities, and men being less likely to engage in help seeking behaviour from a healthcare professional 9.

The majority of studies in this review exclusively used the Fukuda (1994) case definition 2. A common criticisms of this case definition include a combination of broad nonspecific symptoms 1, 36, thus being less likely to identify a homogeneous patient population 1, 36. This may have contributed to a lack of sensitivity of the studies in this review to detect a beneficial effect of the nutritional treatment. Therefore, it is recommended that future research should employ more specific CFS/ME criteria to classify subgroups with similar symptoms and/or severity to identify those patients who may benefit from therapeutic nutrition treatment and counselling.

A variety of instruments, some validated and nonvalidated, were used to evaluate outcome measures in the included studies, thus highlighting a lack of agreement in the best instrument to measure improvement in CFS/ME symptoms, and limiting future comparisons between current and future studies. Therefore, to improve future intervention research, patient outcomes (e.g. fatigue, quality of life) need to be consistently measured using a single validated instrument for each outcome.

Additionally, many of the studies procedures indirectly excluded CFS/ME patients who were house and/or bed bound (i.e. severely affected patients) as a result of their inability to attend a clinic, hospital or university for screening and/or follow‐up. Consequently, sampling bias is likely confounding results, and limits the results to CFS/ME patients with a mild severity of symptoms who have the ability to leave their home. Therefore, future research needs to consider the fluctuating nature and different severities of CFS/ME and thus design flexible protocols that are delivered in a variety of settings (e.g. in clinic or patient's homes) over the telephone or via the post. This will distinguish the effect of different dietary therapies, and identify whether patients with different severity of symptoms respond differently to the intervention.

Nutrition interventions

The aetiology of CFS/ME remains unclear; however, research suggests that this heterogeneous condition is likely a multisystem disorder involving the immune, gastrointestinal, neurological and metabolic systems 1. A majority of the studies in this review investigated the effect of nutritional supplementation to initiate and promote ATP production with respect to reducing patient fatigue and cognitive dysfunction 6, 15, 21, 29. NADH alone 6 and in combination with CoQ10 21 was observed to reduce fatigue in CFS/ME patients. Despite beneficial effects being observed, neither study reported a dietary intake method at baseline or at conclusion of the study to determine whether dietary changes occurred and potentially influenced the results. Furthermore, both studies are limited by small sample sizes and the duration of therapeutic treatment (4 and 8 weeks, respectively). Therefore, longitudinal studies with larger sample sizes are needed to determine whether NADH with and without CoQ10 has a prolonged therapeutic effect on CFS/ME patients. Furthermore, the study by Forsyth et al. 6 is also limited by the use of an investigator‐developed questionnaire to measure symptom outcomes. Despite reproducibility testing being completed, it remains unclear whether this questionnaire was able to identify CFS/ME symptoms. It is recommended that future research uses validated instruments to enable confidence when interpreting results, and to allow comparison between studies.

Ubiquinol‐10 (also known as coenzyme Q10) is an important nutrient for cellular energy production and for its antioxidant function 37. Reduced levels of CoQ10 have been reported in CFS/ME patients compared to healthy controls 38. Only one RCT has investigated supplementation of Ubiquinol‐10 for 12 week in CFS/ME patients 15. Despite no improvement in fatigue, measured with Chandler's Fatigue Scale, Ubiquinol‐10 supplementation improved several other CFS/ME symptoms (e.g. night‐time awakenings), which may have a longer‐term effect on reducing fatigue in CFS/ME patients. To evaluate and demonstrate the benefits of ubiquinol‐10 supplementation, longer‐term controlled studies are needed.

Cocoa and dark chocolate consumption are known to have a number of positive health effect on chronic diseases 39, 40, 41. This review has identified one randomised control cross‐over study that investigated the therapeutic effect of cocoa with respect to decreasing fatigue and improving residual functions, as accessed by the London Handicap Scale, compared to when patients consumed an iso‐caloric low polyphenol chocolate in CFS/ME patients 32. Despite this positive result, the study is limited by the small sample size (n = 10), treatment duration, and a lack of information on dietary habits and intake during the trial. Thus, it is unclear whether energy intake and the micro‐ and macronutrient composition of the diet influenced the result. Furthermore, long‐term observational prospective research and well‐designed RCTs are needed to confirm the clinical effects of cocoa in CFS/ME patients, as well as to understand the mechanism of different types of chocolate in this patient population.

Altered intestinal microbiota may contribute to pathogenesis of CFS/ME 42, 43, 44. It has been postulated that the administration of probiotics may have therapeutic value in CFS/ME patients by decreasing pro‐inflammatory cytokines and improving gut microbiota and mucosal barrier function 45. This review identified two pilot studies each investigating different outcomes. These studies suggest that probiotic bacteria may improve neurocognitive function 33 and anxiety 31. Furthermore, Sullivan et al. 33 observed individual differences regarding improvement in fatigue and physical activity, with patients reporting improvement in one or both areas. This provides further support for stratifying subgroups of CFS/ME patients likely to respond to a therapeutic nutrition treatment. As a result of the individual and preliminary benefits observed in these studies, the future research of probiotics is warranted in a larger homogeneous population.

Previous research suggests that oxidative stress results in CFS/ME patients as a result of diminished antioxidant capacity and/or reduced antioxidant enzymes activity 1, 46. The therapeutic treatment with antioxidants in CFS/ME is limited 20, 25, 30. The studies identified in the present review suggest that the provision of pollen extract may improve CFS/ME symptoms and patients overall wellbeing 25. However, no therapeutic benefits were observed in CFS/ME patients after multinutritional supplementation 20, 30. The results reported by Ockerman 25 are limited by a small sample size (n = 22) and an improvement in certain self‐reported symptoms; thus, the reliability of results remains unclear.

CFS/ME is a complex condition with many symptoms, some of which may be related to the food and beverages consumed. This review identified only one study that evaluated an elimination diet in response to food sensitivities 27. Therefore, future research may consider eliminating potential trigger foods in CFS/ME (e.g. alcohol, caffeine, fat, milk and dairy, gluten), followed by challenges to identify potential problem foods.

Quality assessment

Many of the studies included in this review were of high quality (Rosendal score >60%; Table S1), despite not reporting all methods (e.g. method of blinding, pre‐trial conditions and method for assessing adverse effects). Those studies that received poorer quality scores were pilot or cohort or comparative studies and/or lacked full details of the methods. Future research should describe pretrial conditions and the method for evaluating blinding and accessing adverse effects, as well as the method of dietary assessment at baseline and at the trial conclusion.

Conclusions

Nutrition interventions may be used, in some chronic diseases (e.g. diabetes, cardiovascular disease, obesity), to manage or to minimise the progression of these condition. Therefore, it is reasonable to suggest that nutrition interventions may also improve patient outcomes, such as fatigue and quality of life, in CFS/ME patients. The conclusion of this review supports the current guidelines in that there is insufficient evidence for the use of nutritional supplements and elimination and modified diets with respect to relieving CFS/ME symptoms 47. Therefore, the general prescription of supplements and long‐term elimination diets is not recommended. Rather, recommendations are to eat a balanced diet and a variety of nutritious foods from the basic food groups 47 in accordance with the dietary guidelines for healthy people. However, supplementations may be considered as indicated 47; for example, where CFS/ME patients have diagnoses of irritable bowel syndrome, lactose or gluten intolerance or coeliac disease; or have suspected inadequate nutrient intake; or in cases where nutrient deficiencies are identified via pathology tests.

Studies investigating nutritional interventions in CFS/ME remain very limited, with most interventions being evaluated in a single study. Furthermore, the present review emphasise that the interventions investigated have only been conducted in small sample sizes, and also lacked long‐term follow‐up (>6 months). Despite relative consistency in case definition, the studies differed with regard to inclusion and exclusion criteria and the reporting of participant characteristics (e.g. illness duration, BMI, weight). This heterogeneity in study design presents challenges when aiming to apply findings in the clinical environment. Therefore, longer‐term randomised control trials in homogeneous populations that use more specific criteria are warranted.

Common comorbidities in CFS/ME include FMS and IBS 1. The results of the present study are directly related to CFS/ME only because it did not assess effects on specific comorbid groups. Nutrition interventions targeted toward alleviating symptoms of FMS and IBS, for example, may also reduce the overall functional impact of CFS/ME. Hence, this is an important consideration for future research to inform the dietetic practice, given the high prevalence of comorbidities among CFS/ME patients.

Dietary assessment methods and analysis were not described in any of the studies in the present review. Therefore, dietary variables, other than those being examined, may have confounded the results. The method of dietary assessment for future studies may be derived from a checklist devised by Nelson et al. 48 to adequately describe robust methodological approaches.

Furthermore, to control for dietary variables, it is recommended that future research: (i) assess the intervention for energy and nutrient content and (ii) assess habitual diet at baseline and study conclusion, as well as randomly throughout the trial. This will monitor compliance and minimise diet as a confounding factor in the interpretation of the results. Similarly, to control for physical activity, future research should also consider the use of wearable technology to monitor physical activity levels for the duration of the study.

From a clinical perspective, case studies derived from research with accredited dietitians may also provide additional evidence of benefits to specific cases of CFS/ME, as well as support patients with the dietary regime and fluctuating nature of CFS/ME symptoms.

Transparency declaration

The lead author affirms that this manuscript is an honest, accurate and transparent account of the study being reported, that no important aspects of the study have been omitted and that any discrepancies from the study as planned have been explained. The reporting of this work is compliant with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) guidelines.

The authors declare that they have no conflicts of interest.

The Alison Hunter Memorial Foundation, Change for ME, Mason Foundation, the Stafford Medical Research Foundation, the Edward P Evans Foundation, Queensland Smart State and Advance Queensland provided continued support and funding.

NC designed the study and search strategy, and also performed primary and secondary searches, title and abstract screening, full‐text screening, analysis, quality assessment, and wrote the draft and final manuscript. SJ provided consultation on the study design and data collection and critically reviewed the draft and final manuscript. AC performed the full‐text screening and quality assessment, and also contributed to the final manuscript. DS supervised the study design and critically reviewed and contributed to the final manuscript. SMG supervised the study design and analysis and critically reviewed the manuscript. All authors critically reviewed the manuscript and approved the final version submitted for publication.

Table S1. Methodology quality assessment summary and Rosendal score of studies included in the present systematic review.

Click here for additional data file.