Summary of findings
Description of the condition
HIV/AIDS and macronutrient deficiency
Adequate nutrition is critical for optimal immune function. Dietary therapy is therefore regarded as an important adjunct in the clinical care of patients infected with HIV. It is believed that achieving and maintaining optimal nutrition will improve the individual's immune function, reduce the incidence of complications associated with HIV infection, attenuate the progression of HIV infection, improve quality of life, and ultimately reduce mortality associated with the disease (de Pee 2010; Hsu 2005).
People living with HIV are at great risk of nutritional disorders. This is the case in both untreated (i.e. not receiving antiretroviral therapy (ART)) and treated (i.e. receiving some form of ART) HIV-infected individuals. Furthermore, HIV infection is most prevalent in parts of the world where food security is compromised. Populations at high risk of HIV infection may lack appropriate nourishment prior to infection by HIV. Starvation and undernourishment severely compromise the immune system thereby increasing both susceptibility to HIV infection and progression of HIV/AIDS. (Scrimshaw 1997; Anabwani 2005; Schaible 2007).
Untreated HIV infection is characterised by increased resting energy expenditure (REE, Kosmiski 2011), decreased appetite, decreased intake and digestion of food and decreased absorption of nutrients (Koethe 2010). HIV-infected individuals receiving antiretroviral therapy may experience the adverse effects of antiretroviral drugs such as nausea and insomnia, which also have a negative impact on nutrient status. Poor nutrient status may in turn exacerbate these adverse effects, in part by increased drug toxicity (Ammassari 2001; WHO 2004; Hardon 2007).
As a result of the increased REE in untreated HIV infection, both fat and protein stores are oxidised to fuel the body's energy requirements (Macallan 1995a; Batterham 2001; Kosmiski 2011). Whole-body protein turnover is up to 25% higher in untreated HIV-infected individuals than in HIV-negative controls (Macallan 1995b) and fat stores are replenished more readily than protein stores even when nutrition is adequate (Kotler 1999). The resultant loss of body protein could further compromise the immune system of the HIV-infected individual (Schaible 2007).
HIV-associated weight loss and wasting are independent contributing factors to poor clinical outcomes in people living with HIV/AIDS (Wheeler 1998; Kotler 1989). In most cases, acute weight-loss episodes are associated with secondary infections (Macallan 1993). Once the secondary infections are successfully treated and energy intake is increased sufficiently, patients are able to regain weight and remain weight-stable (Macallan 1998). Instances of chronic weight loss are normally associated with secondary gastrointestinal infections and subsequent malabsorption (Macallan 1993).
Weight loss and low Body Mass Index (BMI < 17kg.m
Description of the intervention
Macronutrients and macronutrient interventions
Macronutrients are variably defined but are classically understood to be essential nutrients that are required by the body in relatively large amounts. Dorland’s Illustrated Medical Dictionary defines macronutrients as carbohydrates, fats and proteins and states that: ‘minerals necessary in relatively large amounts (for example Calcium, Chloride, Magnesium, Phosphorus, Potassium and Sulphur) are sometimes included and sometimes excluded.’ (Dorland 2007).
In this review, a macronutrient intervention could be any intervention given to provide protein and/or energy, ie through carbohydrates and/or fat, by replacing or supplementing the normal diet (for example, high or low fat/carbohydrate/protein diets). Macronutrient interventions can also include dietary supplements not given specifically to provide energy but rather to test the effectiveness of specific nutritional elements (for example, amino acids, whey protein concentrate and Spirulina). Macronutrient interventions may be delivered in liquid, powder or tablet formulation. In resource-scare regions where malnutrition is prevalent, food programmes deliver replacement food or foodstuffs in addition to local staple foods in the form of 1) high-energy ready-to-use therapeutic foods, 2) corn-soya blends, or 3) fortified blended foods, ready-to-use foods, high-energy biscuits and compressed food bars (World Food Programme; Koethe 2009). Macronutrient supplements may be fortified with micronutrients in the form of vitamins and trace elements. An existing Cochrane review, revised in 2010, has summarised the effects of micronutrient supplementation in people with HIV/AIDS (Irlam 2010).
Why it is important to do this review
People living with HIV are at great risk of nutritional disorders. Adequate nutrition is critical for optimal immune function. Dietary therapy is, therefore, an important adjunct in the clinical care of patients infected with HIV.
HIV infection is most prevalent in parts of the world where food security is compromised. Populations at high risk of HIV infection may lack appropriate nourishment prior to infection by HIV. Starvation and undernourishment severely compromises the immune system thereby increasing susceptibility to HIV infection and progression of the disease (Anabwani 2005). Poor nutrient status in HIV-infected individuals is an independent predictor of mortality in both untreated and treated individuals with HIV (Marazzi 2008; Koethe 2010; Liu 2011).
The World Health Organization is currently updating the guidelines for nutritional interventions. Current guidelines for HIV-specific populations are based on an appraisal of the evidence conducted in 2003 and a consultative meeting in 2005. A review of all the evidence conducted on searches as up-to-date as feasible, is desirable when informing guidelines. This 2011 review update seeks to provide a comprehensive summary of the current evidence for macronutrient supplementation in HIV-infected individuals.
The objective of this review was to evaluate the effectiveness of various macronutrient interventions in reducing morbidity and mortality in adults and children living with HIV infection.
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs) evaluating the effectiveness of various macronutrient interventions in the management of individuals living with HIV/AIDS were considered. Studies were included regardless of the setting in which they were carried out.
Types of participants
Studies involving adults and children with HIV/AIDS were included in the review. Studies involving HIV-infected pregnant women were excluded from the review. Studies involving participants with TB/HIV co-infection were included if the randomisation of the participants was stratified according to HIV infection status.
Types of interventions
In this review, a macronutrient intervention could be any intervention given to provide protein and/or energy, ie through carbohydrates and/or fat, by replacing or supplementing the normal diet (for example, high or low fat/carbohydrate/protein diets). Macronutrient interventions could also include dietary supplements not given specifically to provide energy but rather to test the effectiveness of specific nutritional elements (for example, amino acids, whey protein concentrate and Spirulina). In this instance, only interventions providing 1g or more of the specific macronutrient element were included in the review. Only macronutrient interventions administered orally and for four weeks or longer were included in the review.
The control groups could include usual diet, no macronutrient supplementation, dietary counselling or non-nutritive placebo.
Studies assessing the effects of total parenteral or enteral nutritional interventions administered via tube feeding were excluded from this review. Macronutrient interventions fortified with micronutrients were included in the review. However, trials evaluating micronutrient supplements alone in people with HIV infection were excluded, as they form part of an existing Cochrane review (Irlam 2010). Studies assessing nutritional interventions for reducing HAART-related co-morbidities in HIV-infected people were also excluded as these studies would form part of another Cochrane protocol (Marti-Carvajal 2010)
Randomised controlled trials comparing the effectiveness of two or more macronutrient interventions (i.e. studies which did not have a control group that received no macronutrient supplementation, dietary counselling or non-nutritive placebo) in HIV-infected individuals were not included in review, however, data from these studies was included in Table 1.
Types of outcome measures
- All-cause mortality
- Mortality related to HIV infection and other HIV-related conditions
- Morbidity (frequency, types, and duration of episodes of opportunistic infections; incidence of AIDS as defined by each trial; hospital admissions; and other types of illnesses related to HIV infection as reported in each study)
- Disease progression according to WHO (WHO 2007) or CDC staging system as recorded in each study
- Indices of viral load
- Markers of immune response (absolute CD4+ T-lymphocyte count and CD4+ percent of total lymphocytes)
- Nutritional status, including measurements such as body weight, body composition and lean body mass, body mass index (BMI), weight-for height and linear growth in children
- Energy expenditure
- Biochemical markers, such as serum albumin
- Dietary intake and appetite
- Functional outcomes such as child development, quality of life and level of physical activity
If studies reported on additional outcomes not listed above we also reported the data for these outcomes in the review. Adverse events were recorded when these were provided.
Search methods for identification of studies
See: Cochrane HIV/AIDS Group methods used in reviews.
See: Collaborative Review Group search strategy.
A comprehensive, unbiased search strategy was developed to ensure that as many relevant studies as possible were screened for inclusion in the review. An attempt was made to identify all relevant studies, regardless of language or publication status (published or unpublished, in press or in progress).
- Journal and trial databases
A search of MEDLINE was conducted in February 2010 limited from January 2006 to January 2010 as the original review published review in 2007 included searches prior to and including 2006. We ran a final search on 24 August 2011 to ensure that the most recent trial reports were identified using the strategy outlined in Appendix 2: PUBMED search strategy.
The searches were conducted in February 2010 and a final revised search was conducted on the 26 August 2011 using the strategy outlined in Appendix 3: EMBASE search strategy
Cochrane Central Register of Controlled Trials (CENTRAL)
A search of the CENTRAL database of The Cochrane Library (2010), was conducted in February 2010 and a final revised search of Issue 3 The Cochrane Library (2011), on 26 August 2011 using the strategy in Appendix 4: Cochrane Library search strategy. CENTRAL provides records marked either as RCTs or Systematic Reviews.
LILACS ( http://lilacs.bvsalud.org/en/ )
LILACS is an index of scientific and technical literature of Latin America and the Caribbean. We searched this database on 8 February 2012 using the following search terms: (HIV OR HIV/AIDS) AND (nutrition OR nutritional OR nutrient*) [all fields]
- Conference databases
A search of the National Library of Medicine GATEWAY database which contains abstracts from the major HIV/AIDS conferences was done on 5 February 2010. However, no updated search of conference abstracts was done in 2011 as NLM Gateway is no longer maintained and updated.
- Trial Registers
The trials registry www.clinicaltrials.gov was searched on 16 August 2011 using the search string:
(nutrition therapy OR fortified food OR energy intake OR dietary fat OR dietary protein OR dietary carbohydrate OR macronutrient OR spirulina OR supplement OR amino acid OR fatty acid OR oil) AND HIV | Interventional Studies |
The search was limited to 1980 to ensure coverage throughout the period HIV has been documented.
Searching other resources
We were in close contact with individual researchers working in the field, and policymakers based in inter-governmental organizations. We also checked the reference lists of all studies identified by the above methods and examined any systematic reviews, meta-analyses, or clinical guidelines we identified during the search process for references.
Data collection and analysis
Selection of studies
NS and LN read the titles, abstracts and descriptor terms of all downloaded material from the electronic searches to identify potentially eligible reports. Full text articles were obtained for all citations identified as potentially eligible and NS and LN inspected these to establish the relevance of the article according to the pre-specified criteria. Where there was any uncertainty as to the eligibility of the record, the full article was obtained.
NS and LN applied the inclusion criteria with studies reviewed for relevance based on study design, types of participants, interventions and outcome measures. All studies not meeting the inclusion criteria were excluded and the reasons for exclusion were stated in the table “Characteristics of Excluded Studies.”
Data extraction and management
NS, LG, SM and MV independently extracted data using a standardised electronic data extraction form. The following characteristics were extracted from each included trial:
- Administrative details: Trial identification number; author(s); published or unpublished; year of publication; number of studies included in paper; year in which study was conducted; details of other relevant papers cited;
- Details of the study: study design (Cochrane review, non-Cochrane systematic review, RCT); type, duration and completeness of follow-up; country and location of study (e.g. higher-income vs. lower-income country); informed consent and ethics approval;
- Details of participants: presence of malnutrition, age, disease progression according to CD4 count or staging by WHO clinical stages, setting, sample size, relevant baseline characteristics including CD4 count and viral load;
- Details of intervention and control group: type of supplements of macronutrient(s); dosage of macronutrient(s); form and formulation of macronutrient(s), additional co-interventions (such as ART, TB treatment or other management of opportunistic infections); fortification with micronutrients
- Details of outcomes: all pre-specified outcomes and any additional outcomes reported in the study; adverse events and toxicity.
- Details of quality assessment: type of quality assessment including use of the Risk of Bias tool.
- Details of data analysis: numbers and reported statistics for each reported outcome
Where trials were reported in more than one reference, all the trial reports were used to extract data as comprehensively as possible. Discrepancies regarding extracted data were resolved by discussion and, if necessary, referred to an additional reviewer (JV).
Assessment of risk of bias in included studies
The components of each included trial were examined for risk of bias using a standard form. This included information on the sequence generation, allocation concealment, masking (participants, personnel and outcome assessor), incomplete outcome data, selective outcome reporting and other sources of bias. The methodological components of the trials were assessed and classified as adequate, inadequate or unclear as per the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008).
Low risk: investigators described a random component in the sequence generation process such as the use of random number table, coin tossing, cards or envelops shuffling etc
High risk: investigators described a non-random component in the sequence generation process such as the use of odd or even date of birth, algorithm based on the day/date of birth, hospital or clinic record number
Unclear: insufficient information to permit judgment of the sequence generation process
Low risk: participants and the investigators enrolling participants cannot foresee assignment, e.g. central allocation; or sequentially numbered, opaque, sealed envelopes.
High risk: participants and investigators enrolling participants can foresee upcoming assignment, e.g. an open random allocation schedule (e.g. a list of random numbers); or envelopes were unsealed or non-opaque or not sequentially numbered
Unclear: insufficient information to permit judgment of the allocation concealment or the method not described
Low risk: blinding of the participants, key study personnel and outcome assessor, and unlikely that the blinding could have been broken. Or lack of blinding unlikely to introduce bias. No blinding in the situation where non-blinding is not likely to introduce bias.
High risk: no blinding, incomplete blinding and the outcome is likely to be influenced by lack of blinding
Unclear: insufficient information to permit judgment of adequacy or otherwise of the blinding
Incomplete outcome data
Low risk: no missing outcome data, reasons for missing outcome data unlikely to be related to true outcome, or missing outcome data balanced in number across groups
High risk: reason for missing outcome data likely to be related to true outcome, with either imbalance in number across groups or reasons for missing data
Unclear: insufficient reporting of attrition or exclusions
To calculate attrition per group the denominators according to the intention-to-treat principle was used: participants were maintained in the groups into which they were randomized and the overall potential number of participants who could produce an outcome was used as the denominator.
Low risk: a protocol is available which clearly states the primary outcome as the same as in the final trial report
High risk: the primary outcome differs between the protocol and final trial report
Unclear: no trial protocol is available or there is insufficient reporting to determine if selective reporting is present
Other forms of bias
Low risk: there is no evidence of bias from other sources
High risk: there is potential bias present from other sources (e.g. early stopping of trial, fraudulent activity, extreme baseline imbalance or bias related to specific study design)
Unclear: insufficient information to permit judgment of adequacy or otherwise of other forms of bias
Measures of treatment effect
We used Review Manager 5.1 to calculate the risk ratio (RR) for dichotomous data, and the weighted mean difference (WMD) for continuous data, with 95% confidence intervals and we report these results below. Where this was not possible due to missing primary data, we report only the results as presented in the published studies. When units of measurement differed between trials we chose to use the SI unit and appropriate conversions were made.
Dealing with missing data
Where data was missing or unclear, we planned to contact authors where this was possible to do so. When percentages were provided without denominators these were back-calculated. If only standard errors or 95% confidence intervals were reported for means and no standard deviations, the standard deviations were calculated as follows:
SD = SEM x Square root of sample size
Assessment of heterogeneity
RCTs were first assessed for clinical heterogeneity by examining variability in the participants, interventions and outcomes. Where it was clinically meaningful to combine studies, we conducted a meta-analysis using the random effects model as we anticipated heterogeneity. We further assessed statistical heterogeneity in the meta-analysis study results using the Chi-square test for heterogeneity with a 10% level of significance as the cut-off. The impact of statistical heterogeneity was quantified using the I
0% to 40%: might not be important;
30% to 60%: may represent moderate heterogeneity;
50% to 90%: may represent substantial heterogeneity;
75% to 100%: considerable heterogeneity.
Where trials were sufficiently homogenous we combined the results of the trials using the random effects model as we anticipated some heterogeneity. Importantly, for changes in mean measurements, we combined the results if a mean change from baseline to end-point was recorded or if a final mean at the end-point of the study was recorded as described in the Cochrane Review Authors’ Handbook. When measurement time points and lengths of follow up differed between studies we combined these when we thought this was meaningful. Where appropriate we categorised studies into length of follow-up from less than 3 months (short term follow-up), 3 to 6 months (medium term follow-up) and 6-12 months (long term follow-up).
Subgroup analysis and investigation of heterogeneity
The pre-specified sub-groups to investigate for possible sources of heterogeneity across studies were:
- By disease progression: CD4< 200 vs. CD4 <350 vs. CD4 >350
- By disease progression: WHO HIV clinical stage 1-4 or CDC Stages 1-3
- By ART status: on ART vs. not eligible for ART vs. eligible but not on ART
- By age: 0-14 years of age vs. 14-19 years of age vs. 20 years and older
None of the included trials presented data in a format that allowed extraction of the above sub-groups, therefore we did not conduct sub-group analyses.
GRADEPro was used to create Summary of Findings tables for meta-analyses using the pre-specified outcomes. In determining the level of evidence for each outcome, both the efficacy results and the assessment of the risk of bias was integrated into a final assessment of the level of evidence and full details of the decision provided.
Description of studies
Results of the search
2010 Search of electronic journal and trial databases
The 2010 PUBMED search yielded 897 records of which 32 were identified as potentially eligible studies and the full texts obtained.
The 2010 EMBASE search yielded 528 records of which 18 were identified as potentially eligible and the full texts obtained.
The CLIB search yielded 333 records marked as RCT records of which 21 were identified as potentially eligible and the full texts obtained.
2011 Search of electronic journal and trial databases
The 2011 searches of PUBMED yielded 265 records, of EMBASE 152 records and of CLIB 42 records. These were de-duplicated electronically using PROCITE reference management software and 67 duplicate records were removed prior to manual checking.
The NLM GATEWAY search yielded 1503 records of which 89 were coded as Meeting Abstracts. We identified two of these abstract records for further assessment.
The 2011 search of www.clinicaltrials.gov yielded 212 studies of which 28 were identified for further assessment.
One included study was identified through discussion with an expert in the field. A further study was identified through the search conducted for the review entitled: Micronutient supplementation in children and adults with HIV infection. This study has been placed in the Studies awaiting assessment category.
A total of 33 full-text articles were obtained. We conducted eligibility assessments on each of these and identified a total of 6 references reporting on 6 RCTs eligible for inclusion. The reasons for exclusion for 25 studies are detailed in Figure 1. Two studies were categorised as awaiting assessment (see: Studies awaiting classification) and once further details are obtained, these studies may become eligible.
|Figure 1. Study flow diagram.|
We identified a total of six new RCTs to be included in the review in addition to the eight RCTs already included in the review, making a total of 14 RCTs. Full details for each of the trials are reported in the table: Characteristics of included studies.
Seven studies were conducted in high-income countries. Four studies were conducted in the USA (Clark 2000; Keithley 2002; Rabeneck 1998; Shabert 1999), two in Switzerland (Berneis 2000; Karsegard 2004) and one in Germany (Schwenk 1999). Four studies were conducted in African countries. One in South Africa (Rollins 2007), one in Kenya (FANTA-KEMRI study 2011), one in Burkina Faso (Simpore 2005), one in Central African Republic (Yamani 2010). One trial was conducted in Brazil (Moreno 2005) and another in India (Sudarsanam 2011). It is not clear where the study by de Luis 2003 (de Luis 2003) was conducted, although based on the address of the corresponding author it appears to be Spain. In general the trials were conducted in the out-patient (ambulatory) setting.
A total of 1725 adults and 271 children were included across the trials. All participants were infected with HIV. In six trials at least half or all of the participants received antiretroviral therapy (ART). In three of the four African studies none of the participants received ART. In seven studies the majority (>50%) or all participants were CDC category C (AIDS) or WHO Stage III/IV. More men participated in trials than women. Trials evaluated interventions in participants from diverse populations including those with normal weight, those with stable weight loss, and those with malnutrition. In one trial participants were TB/HIV co-infected (Sudarsanam 2011). In the remaining trials all participants were free of confirmed secondary infections or other signs and symptoms of infection, such as fever, chills, or persistent diarrhoea upon enrollment into the study. No trials were identified in pregnant or lactating women although two studies awaiting assessment may provide data on this population group when final results are available (Kindra 2011; van der Horst 2009).
Trial interventions and comparison groups
The interventions can be broadly classified as follows:
Macronutrient supplementation fortified with micronutrients given to provide protein and/or energy by replacing or supplementing usual diet plus nutrition counselling versus nutrition counselling or standard care
Macronutrient formulas fortified with micronutrients plus nutrition counselling versus nutrition counselling alone in participants with and without weight loss (five trials)
- Four trials in participants with > 5% weight loss:
- Liquid supplement containing 600 kcal/day, 17% protein (whey), 59% carbohydrates, 26% fat, electrolytes, trace elements and micronutrients plus nutrition counselling versus nutrition counselling for 12 weeks (Berneis 2000).
- Liquid supplement containing 795 kcal/day, 14% protein, 54% carbohydrates, 32% fat, electrolytes, trace elements and micronutrients (ENSURE
®) plus nutrition counselling versus nutrition counselling for 12 weeks (de Luis 2003).
- Specialized medium chain triglyceride (MCT) formula to provide additional 960 kcal/day over and above usual diet (Lipisorb® Liquid Nutrition: 17% protein, 48% carbohydrates, 35% fat) plus nutrition counselling versus nutrition counselling for 6 weeks (Rabeneck 1998).
- Range of fortified oral supplements providing 600 kcal/day in addition to usual diet plus nutrition counselling versus nutrition counselling for 8 weeks (Schwenk 1999)
- One trial in participants with normal body weight assumed:
- 1-2 cans per day of standard oral formula (Ensure Plus®: 355 calories per can, 15% protein, 53% carbohydrates, 32% fat) plus nutrition counselling versus 1-2 cans per day of immune-enhancing oral formula (Advera®: 303 calories, 19% protein, 65% carbohydrates, 16% fat) plus nutrition counselling versus nutrition counselling for 1 year (Keithley 2002).
Supplementary food fortified with micronutrients plus nutrition counselling vs nutritional counselling alone in malnourished participants (two trials)
- One trial in participants either initiating ART or not yet eligible for ART:
- ART (BMI<18.5 kg/m
2) and pre-ART participants received either 300g per day of fortified blended food for 6 months plus nutrition counselling for 12 months or nutrition counselling alone for 12 months (FANTA-KEMRI study 2011)
- One trial in participants with TB/HIV co-infection:
- A locally prepared cereal-lentil mixture providing 930 kcal/day plus multivitamin micronutrient supplement plus nutrition counselling versus nutrition counselling alone for 6 months (Sudarsanam 2011).
Specific macronutrient supplements versus placebo, no supplement or usual diet
- Two trials in participants with weight loss or low BMI
- Participants received either amino acid mixture containing 14g arginine (free base), 14g glutamine and 3g β-hydroxy-β-methylbutyrate (HMB, calcium salt) plus citric acid (ph 4.5) or isocaloric placebo of bulk maltodextrin for 8 week(Clark 2000)
- ART naive participants received either 10g per day Spirulina or 10g per day placebo of green clay for 6 months (Yamani 2010)
- Two trials in participants with normal or assumed normal weight
- 10g per day of L-ornithine alpha-ketoglutarate (OKG, 1.3g of nitrogen) and nutritional counselling versus isonitrogenous placebo plus nutrition counselling for 12 weeks (Karsegard 2004)
- Daily vitamin and mineral preparation plus L-glutamine (GLN) amino acid (400g/day) plus antioxidant nutrients (ascorbic acid 800mg/d, α-tocopherol 500 IU/d, β-carotene 27000IU/d, selenium 280ug/d and N-acetyl cysteine 2400mg/d) plus weekly nutrition counselling versus placebo of glycine plus weekly nutrition counselling for 12 weeks (Shabert 1999)
Macronutrient supplements fortified with micronutrients given to provide protein and/or energy by replacing or supplementing usual diet versus no supplement or standard care
- One trial in children with prolonged diarrhoea:
Enhanced diet: casein maltodextrin-based milk formula (AL110) until diarrhoea resolved and appetite re-established, thereafter, amount of milk formula modified to provide at least 150 kcal/kg/day containing ˜4.0–5.5 g protein/kg/day and 15% of calories as protein versus standard nutrition support: casein maltodextrin-based milk formula with 67 kcal/100mL offered at least four times per day and a maize porridge/pureed vegetable/oil diet with fermented milk offered at least four times per day. This diet provided at least 100-110 kcal/kg/day containing ˜2.2g protein/kg/day (9.5% of calories as protein) for 6 months (Rollins 2007).
Specific macronutrient supplements versus placebo, no supplement or usual diet
- One trial in children with rapidly progressive HIV infection
Whey protein concentrate versus placebo of maltodextrin for 4 months (Moreno 2005).
- One trial in undernourished children
Spirulina supplement plus traditional meals (millet, vegetables and fruit) versus traditional meals for 8 weeks (Simpore 2005).
None of the studies included in this review specifically aimed to assess the effects of macronutrient interventions on the all-cause or HIV-related mortality and/or morbidity we pre-specified as primary outcomes. Proxies of these outcomes, such as body weight, body composition, viral load, CD4 count, energy intake and quality of life were measured. Outcomes were not consistently reported across trials with some reports conducting analyses with means (normal distribution) and others reporting medians (non-normal distribution) for the same outcomes. In general, outcomes measured anthropometric and immunologic parameters rather than clinical symptoms. Although most trials reported on BMI in the baseline characteristics, few reported on it as an outcome of the trial.
Trials were small with sample sizes ranging from 18 to 1057 with a mean sample size of 143. Follow-up ranged from six weeks to 12 months with more than half (8 out of 14 trials) including a follow-up of 12 weeks or less.
Risk of bias in included studies
Risk due to bias arising from the method of generation of the allocation sequence was low in six studies (Clark 2000; FANTA-KEMRI study 2011; Keithley 2002; Rollins 2007; Shabert 1999; Sudarsanam 2011). Five of the studies reported generating the sequence by computer and for the one study we obtained this information directly from the authors. The remaining nine studies did not report how the randomisation sequence was generated so the risk of bias was rated as unclear in these studies (See: Figure 2; Figure 3).
|Figure 2. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.|
|Figure 3. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
Risk due to bias arising from the method of allocation concealment was rated as low in four studies (Keithley 2002; FANTA-KEMRI study 2011; Schwenk 1999; Sudarsanam 2011). The reported methods included central co-ordination of the allocation and use of sealed, unmarked, opaque envelopes. Allocation concealment was not reported in the remaining ten studies so the risk of bias was rated as unclear for those (See: Figure 2; Figure 3).
Blinding of participants and personnel (performance bias)
Risk of bias arising from lack of blinding of participants and personnel was rated as low for three studies all of which reported using a placebo which was identical in appearance to the intervention (Clark 2000; Karsegard 2004; Shabert 1999). In two trials is was unclear if the participants or personnel were blinded (Moreno 2005; Yamani 2010). The remaining 9 trials all used controls which were not identical to the intervention or the intervention was such that it was not possible to blind the participants and personnel. The risk arising from this was rated as high for all 9 trials (See: Figure 2; Figure 3).
Blinding of outcome assessors (detection bias)
In two trials outcome assessment was clearly reported as blinded, or confirmed with the authors to be so, and detection bias was thus rated as low (Clark 2000; Shabert 1999). Blinding was unclear in 11 trials and the risk of detection bias was rated as unclear. In one trial detection bias was rated as high, as it was stated in the report that no steps were taken to blind the outcome assessors to the treatment assignments (Sudarsanam 2011)(See: Figure 2; Figure 3).
Incomplete outcome data
Risk of attrition bias was rated as low in three trials (de Luis 2003; Simpore 2005; Sudarsanam 2011). Risk of attrition bias was rated as high in nine trials in which attrition was greater than 10% overall and/or differentially distributed between the intervention and control groups (Clark 2000; FANTA-KEMRI study 2011; Karsegard 2004; Keithley 2002; Moreno 2005; Rabeneck 1998; Rollins 2007; Shabert 1999; Yamani 2010). Risk of attrition bias was unclear in the remaining two trials. In Schwenk 1999 overall attrition was moderate at 10%(5/50) but was differentially distributed between groups at 8% (2/26) in the intervention group and 13% (3/24) in the control group (Schwenk 1999). In Berneis 2000 overall attrition was high (16.7%) and no further information regarding group assignment is provided (Berneis 2000)(See: Figure 2; Figure 3).
Effects of interventions
See: Summary of findings for the main comparison Balanced nutritional supplement compared to counselling or nutritional placebo in patients with weight loss for reducing morbidity and mortality in people with HIV
We present the results of the trials conducted in adults followed by the results of the trials conducted in children. Where it was possible to combine the results of outcomes in a meta-analysis, we report on the meta-analysis of these outcomes. When meta-analysis was not possible, we provide reasons for this and report the results for each trial.
Macronutrient interventions, fortified with micronutrients, given to provide protein and/or energy by replacing or supplementing usual diet plus nutrition counselling versus nutrition counselling alone or standard care
MACRONUTRIENT FORMULAS FORTIFIED WITH MICRONUTRIENTS PLUS NUTRITION COUNSELLING VERSUS NUTRITION COUNSELLING ALONE IN PARTICIPANTS WITH AND WITHOUT WEIGHT LOSS
Five trials assessed the effects of macronutrient supplements given to provide protein and/or energy in conjunction with nutrition counselling compared to nutrition counselling alone in adequately nourished participant with or without weight loss (Berneis 2000; de Luis 2003; Keithley 2002; Rabeneck 1998; Schwenk 1999). All the interventions included micronutrients in varied percentages of the Recommended Daily Allowance (details from product web sites). This may be a possible confounder in the outcomes of these studies, as one cannot distinguish between the effect of the increased energy supply or that of the vitamins and minerals. Excluding the Keithley 2002 trial, all the participants entered the trials with a loss of body weight of more than 5% or a reduction in their BMI in the previous 6 months. As Keithley 2002 was a three-armed trial we report the comparisons between each of the two intervention groups and the control group.
- Clinical symptoms
Only Keithley 2002 reported on clinical symptoms but did not define the symptoms nor provide numerical data. The authors report that there were no significant differences between reported clinical symptoms among the two intervention groups and the nutritional counselling control group.
- Energy intake
Meta-analysis of three trials (N = 131; Berneis 2000; de Luis 2003; Schwenk 1999) comparing balanced nutritional supplements (consisting of 50-60% carbohydrate, 15-30% protein and 20-30% fat aimed at improving energy intake by 600-960 kcal/day) with no nutritional supplements, showed significantly increased energy intake by 394kcal/day in the intervention arm (Mean Difference: 393.57kcal/day; 95%CI: 224.66, 562.47; p < 0.00001). See Analysis 1.1. The increase was statistically significant and could be as much as an increase of 562 kcal/day or as little as 225 kcal/day. There was little statistical heterogeneity between the trial results (Ҳ
Although no p values were presented, Keithley 2002 reported no statistically significant differences in energy intake between the groups at any of the study visits. We calculated the mean difference, 95% CI and level of significance for energy intake at 12 months between each group. At 12 months there was no statistically significant difference in the mean energy intake between a) the Ensure Plus group (2236kcal/day, SD 1045) and the control group (1855 kcal/day, SD 991; mean difference: 381 kcal/day; 95%CI: -218.92,980.92; p=0.22); b) the Advera (Immune-enhancing formula) group (2461 kcal/day, SD 1019) and the control group (1855 kcal/day, SD 991; mean difference: 606 kcal/day; 95%CI: -17.30,1229.30; p=0.06); and c) the Ensure Plus group (2236kcal/day, SD 1045) and the Advera group (2461 kcal/day, SD 1019; mean difference:-225; 95%CI:-817.69, 367.69; p=0.46).
Rabeneck 1998 do not provide the numerical energy intake data but report that 56% of the group receiving the specialized MCT formula and 50% of the control group (nutritional counselling) achieved 80% or more of the energy target, defined as 960kcal/day greater than estimated total energy expenditure (p=0.56).
- Protein intake
Meta-analysis of two trials (N = 81; Berneis 2000; de Luis 2003) showed supplementation with a liquid formula providing an additional 600-795 kcal/day and consisting of 14-17% protein, 55-60% carbohydrates and 26-32% fat, significantly increased daily protein intake compared with no supplements (Mean Difference: 23.25g/day; 95% CI: 12.68, 34.01; p < 0.00001). See Analysis 1.2. There was no statistical heterogeneity between results (Ҳ
One trial reported on BMI at baseline and at 6 and 12 months (Keithley 2002). Rabeneck 1998 did not provide data for BMI as an outcome despite recording the method of analysis used for it. In Keithley 2002 mean BMI at 12 months did not differ statistically significantly between the Ensure Plus group (Mean BMI: 24, SD 4 kg/m
- Body weight
A meta-analysis of 4 trials (N=233; Berneis 2000; de Luis 2003; Rabeneck 1998; Schwenk 1999) found no statistically significant difference in body weight between the supplemented group and those receiving nutrition counselling alone (Mean Difference: -0.17; 95% CI: -1.10, 0.75; p = 0.72). See Analysis 1.3. Statistical heterogeneity was quantified at 9% by the I
Keithley 2002 reported no statistically significant differences in body weight between the groups after 12 months, but did not present the p values. Using Review Manager we calculated the mean difference, 95%CI and p values between the groups at 12 months. At 12 months the mean body weight for the Ensure Plus group was 72kg (SD 17) while the mean body weight in the control group was 78kg (SD 23; mean difference: -6.0; 95%CI:-18.23, 6.23; p=0.32). The mean body weight for the Advera group (mean body weight=78kg, SD 13) and the control group (mean body weight=78kg, SD 23) were practically identical (Mean difference was 0, 95%CI: -11.74,11.74; p=1). Even though there was a 6 kg difference in mean body weight between the Ensure Plus group and the Advera group at 12 months this difference was not significant (Mean difference: 6; 95%CI: -2.57, 14.58; p=0.19).
- Fat mass
A meta-analysis of four trials (N = 233; Berneis 2000; de Luis 2003; Rabeneck 1998; Schwenk 1999) showed no statistically significant difference in fat mass, measured as a percentage of total body weight, between the supplemented and non-supplemented groups (mean difference: -1.14%; 95% CI: -2.58, 0.29; p = 0.12). See Analysis 1.4. There was no statistical heterogeneity between results (Ҳ
In Keithley 2002 body fat mass was presented in kg/m
- Fat free mass
In a meta-analysis of three trials (N = 218; de Luis 2003; Rabeneck 1998; Schwenk 1999) there was no statistically significant difference in fat free mass between the supplemented and the non-supplemented groups (MD: -0.37; 95% CI: -2.77, 2.03; p = 0.78). See Analysis 1.5. Statistical heterogeneity between results was not significant (Ҳ
- CD4 cell count
A meta-analysis of the two trials which reported on this outcome (N = 81; Berneis 2000; de Luis 2003) showed no statistically significant difference in CD4 cell count between the supplemented and the non-supplemented groups (Mean Difference: -114.48; 95%CI: -233.20, 4.23; p = 0.06). See Analysis 1.6. Heterogeneity was not significant (Ҳ
Keithley 2002 reported no significant difference in mean CD4 cell counts between the three groups at any of the time points. At 12 months the mean CD4 cell count in the Ensure Plus group was 471 cells/mm
Rabeneck 1998 reports that there was no significant change observed in CD4 counts between the groups (no data provided).
Schwenk 1999 did not report on CD4 counts.
- Viral load
Only one of the five trials (de Luis 2003) reported on the change in viral load over the course of the trial. After 12 weeks no statistically significant difference was noted in HIV viral load (log
- Adverse effects
Adverse effects were poorly reported and in general, were related to tolerance rather than adverse effects. Keithley 2002 reports that no significant differences were found in acceptance and tolerance of the formulas. Rabeneck 1998 noted that one participant discontinued the supplement due to nausea and epigastric pain and one discontinued as he did not like the taste of the supplement.
SUPPLEMENTARY FOOD FORTIFIED WITH MICRONUTRIENTS PLUS NUTRITION COUNSELLING VS NUTRITIONAL COUNSELLING ALONE IN MALNOURISHED PARTICIPANTS
A study conducted in Kenya evaluated the impact of supplementary food on nutritional and clinical status, treatment progress and quality of life of malnourished HIV-infected adults on ART and pre-ART (i.e. HIV-infected adults who did not qualify for ART according to the Kenyan National HIV treatment protocol. ART was only provided to HIV-infected adults who met the WHO Stage IV disease criteria, or who had a CD4 count <200 cells/µl FANTA-KEMRI study 2011). This trial pre-dated the recommendation to commence ART at CD4 counts less than 350 cells/µl. The food products were distributed monthly, for six months or until subjects reach the exit criterion of BMI = 23 kg/m
- Body weight, change in body weight, BMI and % lean body mass
The authors noted that at baseline pre-ART participants had higher weight and BMI compared to the ART participants. Throughout the trial pre-ART participants gained less weight each month than the ART participants.
Amongst participants receiving ART there was no significant difference in mean body weight at any of the time points between the supplement and no supplement group (See Analysis 2.1. However, in the first 3 months of the trial the supplement group appeared to gain weight more rapidly than the no supplement group, as they had a significantly greater change in body weight gain compared to the no supplement group at these time points. After this time point the change in body weight was not significantly different between the groups (See Analysis 2.2). Mean BMI and change in BMI in the supplement group was significantly higher in the first 3 months compared to the no supplement group (See Analysis 2.3). After 3 months there was no significant difference in weight gain, BMI or BMI gain between the supplement and no supplement groups in the participants receiving ART.
Amongst participants not receiving ART, at 3 months (p=0.0027) and 6 months (p=0.001), the supplement group had a significantly greater mean body weight than the no supplement group (See Analysis 2.1). In the first 4 months of the trial and at the 6 month and 9 month time point, the supplement group had a significantly greater body weight gain compared with the no supplement group (See Analysis 2.2). In the first 3 months of the trial and in month 6, 7 and 9, mean BMI in the supplement group was significantly greater than in the no supplement group (See Analysis 2.3). In the first 4 months of the trial and months 6, 7, 8 and 9 the supplement group exhibited a significantly greater gain in BMI than the no supplement group.
There was no significant difference in % lean body mass or changes in % lean body mass between the supplement and no supplement group at any time point for both the ART arm and the pre-ART arm (See Analysis 2.4).
- Clinical and immunological outcomes
There was no significant difference in mean CD4 cell count between the supplement and no supplement group at any time point for both the ART arm and the pre-ART arm. However, in the pre-ART arm, at 3 months, the mean CD4 cell count in the supplement group increased by 7.4±123.5 and the mean CD4 cell count in the no supplement group decreased by 32.59±103.5 (p=0.01, authors own data).
Supplemented ART participants had significantly higher increase in hemoglobin levels at month 3 compared to the no supplement group (change in hemoglobin: 1.23±2.3 in supplement group vs 0.69±2.4 in the no supplement group, p=0.05). Supplemented pre-ART participants had a significantly higher increase in hemoglobin levels at 3 (0.93±2.7 in the supplement group vs 0.01±2.3 in the no supplement group; p=0.01) and 6 months (0.78±3.1 in the supplement group vs -0.18±0.6 in the no supplement group; p=0.05) compared to the no supplement group.
The authors only present the 3 month values for changes in serum albumin. Based on this data there was no difference in changes in serum albumin at this time point between the supplement and no supplement groups for either arms of the study.
- Quality of life
The authors report that most of the changes in quality of life occurred during the initial months of the study. The majority of the ART and pre-ART participants experienced improvements in perceived health or their perceived health remained good, in both supplement and no supplement groups. The authors report that pre-ART participants receiving supplementary food on average experienced significantly greater improvement and less of a decline in perceived health than those not receiving food. These differences did not persist over longer periods of follow-up. Among ART participants, the difference in changes in perceived health between those receiving food and those not receiving food was smaller and not statistically significant (data presented graphically).
In the report, quality of life was also measured by the number of days per month that subjects reported having had poor physical health during the first three months. The authors report that among pre-ART participants, those receiving food had fewer poor health days than those not receiving food, and the differences were significant during the first and second months, but not subsequently. Among ART participants, the difference was significant in the second month only (data presented graphically).
Attrition, defined as discontinuation of care and treatment at the health facility for any reason, including death, loss-to-follow-up, or relocation was high in all groups. Authors report that the highest rate of attrition occurred during the first month of the study: with an attrition rate of 26% in the ART arm and 24% in the pre-ART arm. By the third month of follow-up, 37% and 39% of participants in the food and no-food groups of the ART arm and 37% and 48% of the food and no-food groups in the pre-ART arm respectively had been lost from the study. The difference in attrition between the food and no-food groups in the pre-ART arm was significant (p=.039), and the difference between the groups in the ART arm was not significant (data reported by author).
In the ART arm, the mean durations that participants were retained in the study were 6.3 months and 5.9 months for the food and no-food groups respectively. In the pre-ART arm the mean durations that subjects were retained in the study were 5.9 and 5.4 months for the food and no-food groups respectively (author's results).
The above data is from an unpublished report prepared by the study coordinators (See FANTA-KEMRI study 2011). The report does not include data on all of the measured outcomes, such as adherence to ART, survival, number of severe clinical events (defined as the sum of hospitalizations and deaths) and number of non-severe clinical events (defined as the number of new opportunistic infections and new symptoms for which medication is required).
In a study conducted in India (Sudarsanam 2011) participants with pulmonary tuberculosis, with and without HIV co-infection, and a BMI <19 kg/m² were randomised to receive a cereal-lentil mixture, micronutrient supplement plus standard care versus standard care alone. The supplement consisted of three daily servings of a cereal and lentil mixture (providing 930 kcal and 31.5 g protein) and a once a day multivitamin tablet. Patients were given a months supply of supplement at a time. TB/HIV-coinfected individuals were treated for TB but did not receive ART, as per protocol in India at the time. Despite randomisation the supplemented group had poorer initial nutritional status as measured by most parameters.
There was no significant difference in risk of death at 6 months (22 TB/HIV participants: RR: 2.14, 95% CI: 0.10, 47.38), cure rate at 6 months (22 TB/HIV participants: RR: 1.38, 95% CI: 0.46, 4.14) and treatment failure before 6 months (22 TB/HIV participants: RR: 0.69, 95% CI: 0.12, 4.05) between the two groups. (Data and information retrieved from recently updated Cochrane review: Nutritional supplements for people being treated for active tuberculosis Sinclair 2011).
At the end of the trial there was no significant difference in mean CD4 cell count (See Analysis 5.2: 221±142 cell count.mm
Although supplementation resulted in a significant increase in daily caloric (11.15±882.2 in supplement group vs -375.42±893.2 kcal in no supplement group, p=0.05), protein (4.6±29.2g in supplement group vs -9.85±25.9 g in no supplement group, p=0.019) and fat (2.86±18.9 g in supplement group vs -10.78±17.1 g in no supplement group, p=0.009) intakes, compared to the no supplement group, changes in lean body mass and fat mass in both groups were similar at the end of the trial (change in lean body mass: 2.37±4.97 kg in supplement groups vs 2.40±6.3 kg in no supplement group, p=0.479; change in fat mass: 1.72±4.8 kg in supplement groups vs 1.1±5.4 kg in no supplement group, p=0.573). Data for mean weight are only presented graphically. Authors reported no difference in mean weight changes between the two groups throughout the study. These results are for all the participants, both TB and TB/HIV participants. We have requested information for the TB/HIV participants alone but have not yet received it from the author.
Specific macronutrient supplements versus placebo, no supplements or usual diet
Four trials assessed specific supplements compared either with a placebo or traditional meals in adults (Clark 2000; Karsegard 2004; Shabert 1999; Yamani 2010). Two trials were conducted in participants with weight loss or low BMI (Clark 2000; Yamani 2010). Two trials were conducted in participants with normal or assumed normal weight (Karsegard 2004; Shabert 1999).
ARGININE, GLUTAMINE AND B-HYDROXY-B-METHYLBUTYRATE
Clark 2000 compared the effects of an amino acid mixture containing 14g arginine (free base), 14g glutamine, 3g ß-hydroxy-ß-methylbutyrate (HMB, calcium salt) and citric acid (ph 4.5) with maltodextrin control formulation in participants with unintentional weight loss of 5% or more in the past 3 months who were on ART. After 8 weeks the arginine group gained significantly greater body weight than the control group (MD: 2.63 95% CI: 0.72, 4.54). See Analysis 3.1. There was no significant difference in change in fat mass between the two groups (MD: -0.64 g; 95% CI: -2.69, 1.41; p = 0.54). See Analysis 3.2. The increase in fat-free mass was statistically significantly greater in the arginine group compared with controls (MD: 3.25kg; 95% CI: 1.25 to 5.25; p = 0.001). See Analysis 3.3. CD4 count in the arginine group was reported to increase but not statistically significantly so (p = 0.10). Viral load was reported to decrease statistically significantly (p = 0.007) in the arginine group but actual data for the study endpoint is not presented. Energy intake was not reported on in this study.
The trial by Yamani 2010 was conducted in ARV naive adults from the Central African Republic. A total of 160 patients were included in the trial. Patients in the spirulina group (N=79) received 10g spirulina per day and participants in the placebo group (N=81) received green clay. Each week all participants received food packages from the World Food Program consisting of 14g maize flour, 500g mixture of maize and soya bean, 2kg garden peas, 500g sugar, 150g iodized salt and 500ml oil.
During the 6 month follow-up, 16 patients died and 16 patients were lost to follow up. There was no significant difference in the distribution of lost cases or deaths between the two groups (Spirulina group: 6 deaths and 6 lost to follow-up vs placebo group: 10 deaths and 10 lost to follow-up (p=0.8).
The authors report a statistically significant gain in weight from baseline in each group after 6 months. There was no significant difference in weight gain between the two groups over this time period. Arm circumference also increased significantly in each group, without significant difference between both the groups.
At 3 months the authors reported a significant decrease in mean Karnofsky score in the placebo group versus the spirulina group (p=0.0045). Thereafter the mean Karnofsky score in the placebo group increased to a value similar to that of the spirulina group.
The authors reported a lower number of pneumonia cases in the spirulina group versus the placebo group at 6 months (p=0.01, number of pneumonia cases in each group not presented. Specific information has been requested from author).
After 6 months CD4 count increased in both groups. This increase was not significant and there was no difference in CD4 count between the two groups at the end of the trial. Hemoglobin values decreased in both groups between months 0 and 6, with no significant difference between spirulina and placebo groups.
In both groups serum protein concentration increased in the first 3 months and then decreased in the following 3 months. There was a significant increase in serum protein (g/L) in the spirulina group between months 0-3 (p value not shown) and months 0-6 (p=0.0001). At 3 months (p=0.01) and 6 months (p=0.00001) serum protein concentrations were significantly higher in the spirulina group compared to the placebo group. In both groups, serum creatinine levels decreased at month 3 and then increased again at month 6. At 3 months serum creatinine levels were significantly higher in the spirulina group compared to the placebo group (p=0.01).
MONOHYDRATED L-ORNITHINE ALPHA-KETOGLUTARATE
Karsegard 2004 compared glutarate supplementation with an isonitrogenous placebo containing 9g of milk proteins. Nutrition counselling was provided to all participants. No morbidity data were reported. At study endpoint there were no statistically significant differences between the groups in daily energy intake, body weight, fat mass, fat-free mass or in the CD4 cell counts or viral loads. See Analysis 4.1; Analysis 4.2; Analysis 4.3; Analysis 4.4; Analysis 4.5; Analysis 4.6; Analysis 4.7. The authors conducted ANOVA tests with last known reported values used for missing data and found that there were statistically significant increases in body weight (mean body weight at baseline: 61.2, SD 11.9 in placebo and 56.4 SD 11.2 in OKG group vs mean body weight at 12 weeks: 61.8 SD 11.5 in placebo and 56.8 SD 11.6 in OKG group), BMI (mean BMI at baseline: 20.6, SD 3.0 in placebo and 20.0, SD 2.4 in OKG group vs mean BMI at 12 weeks: 20.8, SD 2.9 in placebo and 20.2, SD 2.5 in OKG group), and triceps skinfold thickness (mean triceps skinfold thickness at baseline: 14.0, SD 5.4 in placebo and 13.4, SD 4.9 in OKG group vs mean triceps skinfold thickness at 12 weeks: 15.5, SD 5.6 in placebo and 15.0, SD 6.1 in OKG group) from baseline in both groups. They do not report specific p values but only p < 0.05 for these outcomes.
The authors report that appetite and food tolerance remained stable over time but that frequency of gastrointestinal events was significantly higher in the OKG group (authors' reported p value < 0.05). Of the 22 participants taking OKG, 19 reported at least one GI adverse event compared with 13 out of 24 in the placebo group with participants in the OKG group one and a half times more likely to experience GI adverse events (RR = 1.59, 95% CI: 1.06, 2.39; p = 0.02).
L-GLUTAMINE AND ANTIOXIDANTS
Shabert 1999 compared a supplement comprising L-glutamine and antioxidants (ascorbic acid, a-tocopherol, ß-carotene, selenium and N-acetyl cysteine) with a placebo supplement comprising 40g glycine. Nutrition counselling and multivitamin supplementation was provided to all participants.
Mean weight at study endpoint was not statistically significantly different between the groups (MD: -1.30kg; 95% CI: -10.18, 7.58; p = 0.77). See Analysis 5.1. Data were not provided to calculate the change in weight over time, but the authors report in the text that the mean change in weight from baseline to week 12 was 2.2kg in the glutamine group and 0.3kg in the placebo group (p =0.04). Similarly mean change in body cell mass (the total mass of all the cellular elements in the body which constitute all the metabolically active tissue of the body; includes muscle tissue, organ tissue, intracellular and extracellular water, and bone tissue) was reported as statistically significantly different between the groups with a greater increase in the glutamine group (mean gain = 1.8kg) than in the placebo (mean gain = 0.4kg). No standard deviations are provided; the authors report a p = 0.007. Mean fat mass was not different between groups at study endpoint (MD: -1.00kg; 95% CI: -32.40, 30.40; p = 0.95). See Analysis 5.2.
Mean CD4 count was not statistically significant between groups (MD: 66.00; 95% CI: -53.39, 185.39; p = 0.28). See Analysis 5.3. The authors conducted an assessment of mood questionnaire at baseline and week 12 using a 30-item profile of the Short Form 36 and report no statistically significant differences between the groups. The authors did not observe any adverse effects in either of the groups.
Three trials were conducted in children. One trial investigated enhanced nutritional support compared with standard nutritional support in children with prolonged diarrhoea (Rollins 2007), one trial investigated a whey protein concentrate compared to maltodextrin placebo in children with rapidly progressing HIV (Moreno 2005) and one trial investigated the effects of spirulina plus traditional meals versus traditional meals alone in malnourished HIV-infected children (Simpore 2005).
Macronutrient supplementation (fortified with micronutrients) versus no supplement or standard care
ENHANCED NUTRITIONAL SUPPORT VS STANDARD NUTRITIONAL SUPPORT IN CHILDREN WITH PROLONGED DIARRHOEA
The primary outcome of the study by Rollins 2007 was weight change from enrolment until 8 weeks among 169 HIV-infected children. The study reported that children receiving enhanced nutrition support had significantly more weight gain in the first 8 weeks than children receiving standard care (weight gain was expressed as the median change in age- and sex specific weight standard deviation scores (SDS): +1.02 (enhanced nutritional support group) vs +0.01 (standard nutritional support group), p < 0.0001).
Mean weight-for-age standard deviation score (WFA-SDS) at baseline of those who dropped out was not different between groups: (Standard nutritional support group dropouts: –3.24 vs; enhanced nutritional support group dropouts: –3.39; p > 0.05) standard nutritional support group completers: –3.00, vs enhanced nutritional support group completers: –3.19, p > 0.05). Therefore, the observed outcome among the completers appeared to be unaffected by dropout bias.
After 8 weeks, weight gain was similar between groups, with only a slight increase in weight-SDS over time in each group. Median change in weight-SDS from 8 to 14 weeks of follow-up: +0.21 (enhanced nutritional support group) vs. 0.00 (standard nutrition support group: 14 to 26 weeks of follow-up: +0.12 (enhanced nutritional support group) vs. +0.24 (standard nutritional support).
Children receiving enhanced nutritional support exhibited significantly higher median attained WFA-SDS after 8 weeks (-1.99 vs -3.06, p<0.05), 14 weeks (-1.69 vs -2.96, p,0.05) and 26 weeks (-1.32 vs -2.63, p,0.05) compared to the children receiving standard nutritional support. Children in the enhanced nutritional support group who were randomised to continue with the enhanced nutrition support until 6 months showed a further improvement in median attained WFA-SDS (−1.01 SDS) at 6 months compared to those who reverted back to standard home diet who maintained the same median value attained at 3 months (-1.68 SDS).
No significant differences in CD4 cell counts or viral load between groups or over time were reported. However, it should be noted that changes in viral load could only be assessed in a subset of 70 children. The pattern of morbidity throughout the study was also similar between groups based on the cumulative frequency of clinical signs recorded during the follow-up intervals of 0–8, 8–14 and 14–26 weeks.
Dropout rates at the end of the study were high in both groups (Standard nutritional support: 27/83 (33%); Enhanced nutritional support: 38/86 (44%); difference 11%, 95%CI: −3 to +26). Authors reported the primary reason for attrition in the study was death. There was no significant difference between the two groups for death at 8 weeks (Death at 8 weeks: 10/83 (12%) in standard nutritional support group vs 14/86 (16%) in enhanced nutritional support group; OR: 1.42 95% CI: 0.59, 3.40; p=0.43; See Analysis 6.1) or death at 26 weeks (Death at 26 weeks: 18/83 (22%) in standard nutritional support group vs 25/86 (29%) in enhanced nutritional support group; OR: 1.48 95% CI: 0.74, 2.98; p=0.27; See Analysis 6.2).
Specific macronutrient supplements versus placebo, no supplements or usual diet (two trials)
Two trials assessed supplementation with specific macronutrients compared either with a placebo or traditional meals in children. One trial was conducted in Brazilian children with rapidly progressive HIV infection (Moreno 2005). One trial, conducted in Africa investigated the effects of Spirulina in undernourished children (Simpore 2005).
WHEY PROTEIN CONCENTRATE
Moreno 2005 was conducted in Brazil in rapidly progressive HIV vertically infected children (N=18). Participants received whey protein concentrate, maltodextrin or placebo. In the final data analysis the results of the maltodextrin and the placebo groups were combined (will be referred to as the placebo group) and compared with those of the whey protein concentrate group. All children were on some form of ART.
Overall attrition was high at 27.7% (5/18), but similar in both groups (3 lost from whey protein concentrate group vs 2 lost from the placebo group). Authors reported that 22.2% (2/9) of the children in the whey protein concentrate group developed co-infections during the 16 week study compared to 77.7% (7/9) of the children in the placebo group (p=0.0567, author's data).
While a non-significant increase in median CD4 cell count was demonstrated in both groups, no significant difference in median CD4 cell count between the two groups was noted throughout the study. Median CD8 cell count decreased over time in the whey protein concentrate group, with a significant difference noted between the 8 weeks and the 16 week values (p=0.046). Since viral load was only measured in 6 children in the whey protein concentrate group, differences between treatment groups could not be assessed. There was no significant difference between the groups for levels of leukocytes, erythrocytes, hemoglobin and platelets.
The trial by Simpore 2005 was conducted in undernourished HIV-negative and HIV-positive children in Burkina Faso. Only the results of 84 HIV-positive children who participated in the trial will be reported in this review. The intervention group received 10g spirulina daily along with traditional meals of millet, vegetables and fruit whereas the comparison group received traditional meals only for 8 weeks.
There was a significant increase in weight-for-height z scores (WHZ) in both groups after 8 weeks. There was no difference in WHZ between the groups at the end of the study. See Analysis 7.1. There was also a significant increase in weight-for-age z-scores in both groups after 8 weeks but no differences between the groups were noted. See Analysis 7.2.
Authors reported that treatment compliance was excellent and none of the children dropped out or were lost to follow up.
Summary of main results
Our review establishes that experimental evidence on the effects of macronutrient supplementation on important clinical outcomes in HIV-infected adults and children is surprisingly limited; despite an exhaustive search, we found only 14 relatively small randomised controlled trials that met the inclusion criteria of our review. Apart from the two trials evaluating Spirulina, one trial in adults and one trial in children, each of the trials evaluated a different macronutrient supplement. Participants included in the trials also varied quite considerably in terms of stage of HIV, HIV treatment status and general nutrient status. Only three trials reported on mortality, two trials in adults and one trial in children. Meta-analysis was only possible on one class of supplements, viz. balanced macronutrient supplements fortified with micronutrients plus nutrition counselling versus nutritional counselling alone or standard care. GRADE assessments were conducted on outcomes for this meta-analysis and included reviewing the data and the potential biases in each trial before grading the level of evidence. None of the trials were graded as providing strong evidence primarily because of small sample sizes (even after meta-analysis) and the high risk of bias due to a lack of blinding, high attrition and unclear reporting of randomisation methods.
In adult participants with weight loss, balanced macronutrient supplements aimed at improving energy intake by 600-960 kcal/day increased intakes of energy and protein compared with no supplement or nutrition counselling alone, but had no effect on other anthropometric or immunologic parameters. This is an important finding as loss of appetite and consequent decreased food intake is common in people with HIV. The increased energy and protein intake in the supplemented group resulted in increased body weight in most cases. However, only in de Luis 2003 was the increase in body weight in the supplemented group significantly greater than that of the group receiving nutrition counselling alone.
As all the supplements included micronutrients supplementation, it is not possible to determine if the observed effects of an increase in energy and protein intake are due to the combination of the macro- and micronutrient components or the macronutrient component alone. Variation in the nutritional composition of active supplements and in control interventions across studies, as well as differences in the disease stage of the participants in various studies (which can significantly affect intake and absorption of food) limits the value of meta-analysis and complicates the interpretation of the findings of this review. All the trials were of a 12 week or less duration making inferences prone to inaccuracies.
It is worth noting that patients with acute opportunistic infections are most prone to weight loss and theoretically this sub-group would be most likely to experience improvements in nutritional status such as weight gain. However, this sub-group was excluded from all of the studies evaluating balanced macronutrient supplements. Nonetheless, if the finding of increased energy and protein intake in our review is valid, this would be important, given that decreased energy intake, increased resting energy expenditure and accelerated protein turnover is common in people with HIV/AIDS.
In contrast to studies in adults with weight loss, balanced macronutrient supplementation did not have any effect on food intake, nutritional status or anthropometry in adults with assumed normal weight (based on results of one small trial, N=66).
A relatively large trial (N=1057) conducted in Kenya evaluated the effect of providing supplementary food to malnourished adults either initiating ART or pre-ART. Overall, malnourished HIV-infected participants not yet receiving ART (pre-ART) gained less weight than those receiving ART, regardless of supplement status (i.e. whether receiving supplementary food or not). Although there was no significant difference in body weight in the first three months of initiating ART between the supplemented and non-supplemented groups in the ART arm of the trial, participants receiving supplementary food showed a significantly greater weight gain and BMI gain (i.e. the change in weight and BMI at each time point in the first 3 months of the trial was significantly greater in the supplemented group compared to the son-supplemented group)than those not receiving supplementary food. This was evidenced by the significantly greater changes in body weight and BMI at these time points in the supplemented group of the ART arm. These effects did not persist after the first 3 months of the trial. The beneficial effects of supplementary food on body weight, body weight gain, BMI and BMI gain was more pronounced and more prolonged in the pre-ART participants than in the ART participants. At almost every time point in the first 9 months of the trial, pre-ART participants receiving supplementary food weighed significantly more and had a significantly higher BMI than the non-supplemented participants. Supplementary feeding did not affect CD4 cell count or serum albumin levels in either ART or pre-ART participants. Supplementary feeding had a significant beneficial effect on haemoglobin levels in both ART and pre-ART participants. In the initial stages of the trial supplementary food had a significant beneficial effect on the quality of life of pre-ART participants in particular.
In India, providing supplementary food to TB/HIV co-infected adult participants not yet receiving ART did not significantly alter risk of death at 6 months, TB cure rate at 6 months or TB treatment failure before 6 months compared to TB/HIV co-infected participants not receiving supplementary food. The provision of supplementary food did not significantly alter CD4 cell counts or viral load compared with values from the non-supplemented participants.
HIV infection is most prevalent in parts of the world where food security is compromised. These populations at high risk of HIV infection lack appropriate nourishment prior to infection by HIV. Poor nutrient status in HIV-infected individuals is an independent predictor of mortality in both untreated and treated individuals with HIV (Marazzi 2008; Koethe 2010; Liu 2011). These two trials evaluating supplementary feeding in low-income countries are important as they provide some evidence of the effectiveness of macronutrient supplementation in these populations who are likely to benefit the most from such interventions.
Four small trials evaluated four different specific macronutrient interventions so meta-analysis was not possible. In adult participants with unintentional weight loss of 5% or more supplementation with an amino acid mixture containing arginine, glutamine and beta-hydroxy-beta-methylbutyrate (HMB) significantly increased body weight and fat free mass and significantly reduced viral load. Various anthropometric and immunological outcomes were not significantly affected by daily supplementation with 10g of spirulina compared to placebo of green clay in malnourished, antiretroviral naive participants. It was reported that participants receiving spirulina had a significantly higher Karnofsky score and significantly fewer cases of pneumonia compared to the participants receiving placebo.
In adult participants with assumed normal weight, supplementation with monohydrated L-ornithine alpha-ketoglutarate (OKG) did not significantly alter anthropometric or immunological outcomes compared with a placebo as both groups showed similar increases in all of the measured indices throughout the study. Although there was no significant difference in mean body weight between the groups at the end of the study, supplementation with l-glutamine and antioxidants resulted in a significantly greater mean weight gain over 12 weeks compared with placebo.
Three small trials were conducted in children. All of the trials were conducted in low- to middle-income countries, namely South Africa, Brazil and Burkina Faso. In children with prolonged diarrhea enhanced nutritional support (providing an extra 50kcal/kg/day, double the amount of protein and 5% more protein as a percentage of total calories than standard nutritional support) resulted in significantly greater weight gain after 8 weeks compared with the standard nutritional support. Attrition was high in both groups, mainly due to death of participants. It is worth noting that 43 children died during the 6 month trial. None of these children were receiving ART as was standard care in South Africa at the time of the trial.
The only notable result from the small trial evaluating the effects of whey protein concentrate in rapidly progressive HIV vertically-infected children was the significant reduction in the number of co-infections experienced by the children receiving the whey protein concentrate, compared to those receiving placebo. While this is a very important finding it should be interpreted with caution as this was a very small trial of 18 children and attrition was high in both groups (33% in whey protein concentrate group and 22% in the control group). All of the children were on some form of ART. In undernourished HIV-infected children the addition of 10g of spirulina to daily traditional meals did not significantly alter weight-for-height or weight-for-age z scores compared to the children not receiving spirulina. Compliance was excellent and there was no attrition.
Overall completeness and applicability of evidence
Only two of the 14 trials included in this review evaluated the same supplements. Of those trials that evaluated similar classes of supplements, none of the supplements were identical in composition, formulation or quantity. Comparisons and control interventions also differed across every trial. The length of follow-up varied from 6 weeks to 12 months and few trials measured the same suite of outcomes. Due to this high degree of variation, it was not possible to conduct meta-analysis on most of the trials, leading to reduced confidence in the results arising from single trials. Stronger evidence would be gained from replicating trials of each of the supplements evaluated. This would also facilitate meta-analysis.
Earlier trials evaluated the effects of supplements on body weight and other measures of body composition, but few report on outcomes related to morbidity or mortality. While mortality and morbidity were not the primary outcomes measures of the more recent trials, the majority of them do report on these important outcomes.
Seven of the eight trials included in the original version of this review were conducted in high-income countries, in relatively well nourished (in terms of body mass index) adult males and females between the ages of 30 and 50, almost all of whom were on some type of ART. It is promising to note that all of the six new trials added to the review were conducted in low-to middle-income countries (four trial in Africa) and the interventions tested in these trials are relevant and accessible to the populations in these countries. The addition of three studies conducted in children is also noteworthy. There are major differences between HIV-infected individuals in rich and poor countries with regard to income level, availability of adequate nutrition, and access to basic healthcare or life-prolonging antiretroviral treatment. In addition, patient groups vary in terms of their response to the virus, stage of the disease, susceptibility and exposure to opportunistic infections, nutritional status, and individual response to the various treatments received. These observations highlight some challenges in applying the findings of trials conducted in high-income countries to people in resource-constrained countries, where nutritional status is generally poor and access to antiretroviral therapy is limited. In other words, where supplementation may be most needed.
Importantly this review identified only two potentially eligible trials (Kindra 2011; van der Horst 2009) which were conducted in pregnant and lactating women infected with HIV. Under normal conditions, successful pregnancy and lactation requires an increase in energy intake. In HIV-infected women, HIV-associated anaemia can be exacerbated during pregnancy. This review has identified that we do not have evidence from trials in support of macronutrient supplementation (either fortified with micronutrients, such as iron, or not) in this population.
Table 1 provides information on RCTs comparing two or more macronutrient interventions in HIV-positive individuals. While we did not include the results of these trials in our review we have provided this information for the sake of completeness.
Quality of the evidence
GRADEPro was used to create Summary of Findings tables for meta-analyses using the outcomes pre-specified by the WHO expert group. In determining the level of evidence for each outcome, both the efficacy results and the assessment of the risk of bias was integrated into a final assessment of the level of evidence and full details of the decision provided in footnotes. All primary and secondary outcomes were rated as critical by the WHO expert group and received a score of ‘9’ indicating the highest level of importance to patients. For other outcomes included in the meta-analysis, levels were determined at discretion of the authors.
Balanced supplementation versus counselling
See Figure 4 for complete assessment and rationale for ratings.
|Figure 4. GRADE Summary of Findings Table for balanced macronutrient supplements versus counselling|
For the outcomes of an increase in the mean change in energy and protein intake the level of evidence was rated as low.
For body weight, and fat mass measured in percentage of total body weight, the level of evidence was rated as moderate.
Evidence for fat-free mass and change in CD4 count from baseline to study endpoint was rated as low.
The evidence for an effect on viral load was rated as very low with this result coming from only one trial.
Potential biases in the review process
In the meta-analysis of trials evaluating balanced macronutrient supplements fortified with micronutrients and nutrition counselling versus nutrition counselling we combined data reflecting both mean change from baseline to end-point and final mean at the end-point of the study as described in the Cochrane Review Authors’ Handbook. Caution must be used when interpreting the outcomes of this meta-analysis as certain effects of the interventions may be hidden when combining these two types of data. Furthermore, the lengths of follow up of the trials included in this meta-analysis were also different (two trials had a 6 week follow up, one trial had an 8 week follow up and one trial had a 12 week follow up). This could also introduce bias into the results of the meta-analysis. Finally, the interventions evaluated in the trials were similar, in a very broad sense, but not identical.
Biases in the review process were minimised by performing a comprehensive search of the literature, independently selecting and appraising the studies, and extracting the data in duplicate. Where data was missing, we sought additional information and data directly from authors where this was possible to do so. We also consulted product web sites to gain more detailed product information regarding several of the commercial supplements.
Although an extensive hand-search for grey literature was not conducted, it is unlikely that important trials have been missed given the high profile nature of the topic and the close partnership established with agencies and organizations working in this area. Indeed, due to the involvement of the WHO in this review, the authors were presented with data from the unpublished FANTA-KEMRI trial. However, the review remains at risk of publication bias from less prominent trials. We attempted to reduce this risk by identifying relevant conference abstracts. A 2010 search of the HIV/AIDS-specific conference database, NLM Gateway, yielded two relevant trials, both of which were already included in the review. Given that NLM Gateway is no longer maintained, conducting an additional search of the new HIV/AIDS conference database, www.aegis.com, could potentially yield unpublished studies and is warranted to reduce the possible impact of publication in this review.
The search of the trials registry, www.clinicaltrials.gov, to identify trial protocols and ongoing trials yielded 28 potentially relevant trials protocols. These will require further assessment and exploration to either 1) link them to trials already included in the review, or 2) if not included, to attempt to obtain the completed trial reports. This task is time-consuming and has to be balanced against feasibility and time constraints.
Implications for practice
Implications for research
In summary, there is limited evidence, from randomized trials mainly conducted in high-income countries that targeted supplementation of the diet with balanced macronutrient supplementation increases energy and protein intake in HIV-infected adults on antiretroviral therapy. The effects of nutritional supplementation on mortality, morbidity, body weight, and immunological parameters remain unclear. There is limited evidence for the optimal type and amount of supplement delivered by food programme interventions for people infected with HIV.
The authors would like to thank Este Vorster, Ulrich Keller, Douglas Wilmore, Claude Pichard, John Rathmacher, Achim Schwenk, Joyce Keithley, Elizabeth van der Merwe, and Wieland Gevers for methodological input or assistance in the interpretation of data. The initial review was completed with the guidance and support of the HIV/AIDS mentorship programme coordinated by the South African Cochrane Centre and the HIV/AIDS Review Group. Sarah Mahlungulu received a bursary from the Cochrane Health Promotion and Public Health Field to conduct the review. For the 2011 revision the support of the World Health Organization Nutrition for Health and Development Department, Switzerland.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Index terms
Appendix 1. World Health Organization clinical staging of HIV/AIDS for adults and adolescents with confirmed HIV infection
Appendix 2. PubMed search strategy
Appendix 3. EMBASE search strategy
Appendix 4. Cochrane Library search strategy
Last assessed as up-to-date: 22 May 2012.
Protocol first published: Issue 4, 2003
Review first published: Issue 3, 2007
Contributions of authors
SM developed the protocol for this review with the help of MVIS and JV. LG assisted SM and MVIS with data extraction and went on to complete the analysis and the write up of the final review. SM, MVIS and JV assisted in editing the final review. SM, LG and NS screened trials for inclusion in the review. LG and NS conducted the data extraction, analysis and write up of the updated review. JV, MVIS and SM assisted with editing the final version of the updated review.
Declarations of interest
We declare that we have no affiliation with or involvement in any organisation or entity with a direct financial interest in the subject matter of review (e.g. employment, consultancy, stock ownership, honoraria, or expert testimony).
Sources of support
- UNIVERSITY OF NATAL, DURBAN, South Africa.
- SOUTH AFRICAN COCHRANE CENTER, South Africa.
- COCHRANE HEALTH PROMOTION AND PUBLIC HEALTH FIELD, Australia.
- WHO Nutrition for Health and Development Department, Switzerland.
Medical Subject Headings (MeSH)
Diet [standards]; Dietary Carbohydrates [*administration & dosage]; Dietary Fats [*administration & dosage]; Disease Progression; Energy Intake; HIV Infections [complications; *diet therapy; mortality]; Randomized Controlled Trials as Topic
MeSH check words
Adult; Child; Humans
* Indicates the major publication for the study