Summary of findings
This review will not be updated. It should be considered out of date. It has been superseded by this review:
Irlam JH, Siegfried N, Visser ME, Rollins NC. Micronutrient supplementation for children with HIV infection. Cochrane Database of Systematic Reviews 2013, Issue 10. Art. No.: CD010666. DOI: 10.1002/14651858.CD010666. Link to Cochrane Library. [PubMed]
Haematologic abnormalities are common in children with HIV. The most prevalent haematologic disturbance associated with HIV in children (apart from CD4 lymphocytopenia) is anaemia (Ellaurie 1990; Forsyth 1996; Eley 2002; Adetifa 2006). It is a associated with more-rapid progression to AIDS and decreased survival rates (Forsyth 1996; Moore 1999). The prevalence of anaemia in paediatric HIV ranges from 23% to 48% in high income countries (Pizzo 1988; Butler 1991; Forsyth 1996) to 78% to 90% in low income countries (Adewuyi 1994; Totin 2002; Clark 2002; Adetifa 2006). Anaemia associated with HIV can arise from multiple mechanisms, including direct infection of haematopoietic precursor cells (Zauli 1994; Carr 1998), malaria, opportunistic infections, micronutrient deficiency, anaemia of chronic disease (ACD), and direct effects of medications, such as the antiretroviral zidovudine, as well as other drugs given for prophylaxis and treatment of opportunistic infections (Richman 1987; Pizzo 1988; Semba 2001; Totin 2002). Anaemia of chronic disease is the most common type of anaemia and is due to the inhibitory effect of inflammatory cytokines released during chronic diseases on erythropoiesis. These cytokines also blunt the response to erythropoietin, prevent the release of iron from the reticuloendothelial stores, and shorten red cell survival time (Coyle 1997; Allen 1998).
Iron deficiency is the most common cause of anaemia of micronutrient-related aetiology worldwide, affecting about 50% of women and children in developing countries and about 25% of men (WHO 2001). There is considerable geographic overlap of areas of the world where both iron deficiency anaemia (IDA) and paediatric HIV are distributed. These areas include sub-Saharan Africa and Southeast Asia, where iron supplementation is routinely given to anaemic children and pregnant women. Given this high prevalence of IDA, it is likely that many HIV-infected children also are iron deficient. The contribution of iron deficiency to the anaemia of HIV in children has been described but is not completely understood. Iron deficiency, iron depletion, and iron-deficient erythropoiesis have been found in over 50% of anaemic HIV-infected children. This finding is positively correlated with the severity of clinical disease and immunologic suppression (Eley 2002; Totin 2002). Iron status usually is determined by measuring haemoglobin concentration and red cell indices; microscopically examining red cell morphology, plasma ferritin, transferrin, and total iron binding capacity (TIBC); and directly determining levels of iron. Distinguishing ACD from IDA as the underlying cause of anaemia in HIV can be difficult because many of the direct and surrogate markers of iron status behave like acute phase reactants and may be elevated in chronic inflammatory diseases, even in the presence of iron deficiency (Feelders 1999). Elevated levels of ferritin and transferrin have been found in iron-deficient HIV-infected children with profound immunosuppression and viral burden and advanced clinical stage of illness (Eley 2002; Totin 2002). It is therefore logical to contemplate supplemental or therapeutic courses of iron in HIV-infected children. At present there are no guidelines for the prevention and management of anaemia in HIV-infected persons, including children (Volberding 2000).
There are reasonable hypothetical considerations and in vitro experiments that suggest that excessive iron may have deleterious effects in persons living with HIV. These effects include accelerated progress to AIDS with increased viral load and falling CD4 counts, impaired innate and specific immune responses, increased oxidative stress, and reduced survival rates (Boelaert 1996; Savarino 1999). This evidence is based largely on several observational studies in HIV-infected adults from populations with little or no prevalent iron deficiency (Costagliola 1994; Salmon-Ceron 1995; Delanghe 1998; De Monye 1999). The survival analysis was not adjusted for baseline clinical stage of disease, degree of immunosuppression, and viral burden of the individual subjects in any of these reports and studies. In contrast, another study reports that low-dose iron supplementation did not increase HIV-1 load (Olsen 2004).
Currently, iron supplementation for anaemic infants and children is routinely practiced without any obvious effect in developing countries, which bear most of the burden of global paediatric HIV infections.
Because iron deficiency and IDA are common in HIV-infected children in high-prevalence areas and because there are concerns over a possible deleterious effect of iron, this review aims to assess the evidence for iron supplementation for reducing morbidity and mortality in HIV-infected children.
To determine the safety and efficacy of iron supplementation in improving clinical, immunologic, and virologic outcomes in children infected with HIV
Criteria for considering studies for this review
Types of studies
All randomized controlled clinical trials that compare iron supplementation with other products and/or placebo.
Types of participants
Children aged 12 years and below with human immunodeficiency virus infection, from hospital or out-patient (ambulatory) settings. Children may be receiving HAART.
All studies of children with bleeding disorders will be excluded.
Types of interventions
Iron supplementation in any form, dosage and administered by any route for all durations
Types of outcome measures
- Progression to AIDS,measured by the incidence of AIDS defining illnesses (clinical definition) and by CD4% and/or absolute CD4 counts (immunologic definition).
- Change in CD4 counts, measured as the mean of absolute changes over time and change across immune categories.
- Change in viral load over time will be measured in logs. A 2 log change in viral load will be considered significant between measured periods.
- Number of hospital admissions will be measured as number of admissions over period of follow-up.
- Duration of hospitalization. The mean duration will be measured.
- Need for blood transfusions will be recorded as the number of transfusions needed and mean volume of blood required during study period.
Search methods for identification of studies
See: Cochrane HIV/AIDS Group methods used in reviews.
See: HIV/AIDS Collaborative Review Group search strategy.
With the assistance of the HIV/AIDS Review Group Trials Search co-coordinator, we formulated a comprehensive and exhaustive search strategy in an attempt to identify all relevant studies regardless of language or publication status (published, unpublished, in press, and in progress). Full details of the Cochrane HIV/AIDS Review Group methods and the journals handsearched are published in The Cochrane Library in the section on Cochrane Collaborative Groups (http://www.mrw.interscience.wiley.com/cochrane/clabout/articles/HIV/frame.html).
We used the RCT search strategy developed by The Cochrane Collaboration and detailed in the Cochrane Handbook for Systematic Reviews of Interventions in combination with terms specific to iron deficiency.
The following electronic databases were searched
- MEDLINE (1966 to date) via PUBMED in September 2007 and updated in November 2007 using the strategy described in Table 1
This search yielded 89 references from which we selected three for full article retrieval.
- EMBASE (1980 to 2007) in December 2007 using the strategy described in Table 2. This search yielded 38 references from which we selected two for full text retrieval.
- We also searched NLM Gateway and AIDSearch, which includes coverage of the following conferences:
-International AIDS Conferences (1985 to date)
-Conference on Retroviruses and Opportunistic Infections (1986 to date)
-The British HIV Association conference (1997 to date)
-International Congress on Drug Therapy in HIV infection (1994 to date)
- Searching AIDSearch (1980-2007) in December 2007 using the search strategy described in Table 3 yielded 60 records, of which we selected three for full article retrieval. Searching the NLM Gateway (1980-2007) in December 2007 using the strategy shown in Table 4 yielded 13 records, but none were selected for full article retrieval.
- The Cochrane Controlled Trials Register, published in The Cochrane Library (the latest issue available at the time) using the search strategy described in Table 5. This contains mainly reference information specific to randomised control trials and controlled clinical trials in healthcare. This yielded 6 references in total of which we selected 2 for full article review.
There was some overlap between references retrieved in each database. We also checked the citations of all the trials identified by our search strategy for relevant studies.
The clinical registers (ClinicalTrials.gov and Current Controlled Trials) were also searched. No additional ongoing studies were found.
Data collection and analysis
Reviewers IMOA and UO independently screened and selected abstracts of all records identified for eligibility. Description of interventions, study design, available outcomes, and clinical data were extracted from the abstract using a brief, standardised eligibility form. Studies were rejected on initial screening if they were not RCTs with at least one comparison group of non-supplemented children.
Agreement between the two reviewers was reached by discussion with the HIV/AIDS Mentor (NS) acting as final arbiter in case of disagreement. Full papers were then obtained for all abstracts passing the initial screening and for records in which eligibility was uncertain. These papers were then assessed for eligibility in more detail and classified according to the review objectives.
The major headings in the standardised data extraction form included the following.
- Administrative details: identification; authors; 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 study: study design;type, duration and completeness of follow-up; country and location of study (developed versus developing country).
- Characteristics of participants: prior exposure to antiretroviral therapy (type and duration); disease stage; baseline CD4 count; baseline HIV RNA level.
- Details of intervention: types and doses of drugs used; duration of therapy; adverse events (methods of surveillance for adverse events and type: minor or severe) and adherence measures (if reported).
No RCTs met the inclusion criteria. The reasons for exclusion of retrieved full text articles are given in the table of excluded studies. However, should any trials meet the inclusion criteria in the future, we will assess quality and extract and analyse data as described in this section.
We planned to qualitatively capture trial methods by recording the method of generating the randomisation sequence; the adequacy of allocation concealment; the adequacy of blinding in the participants, providers and assessors; and the differential loss-to-follow-up in the comparison groups. Had we found any trial, we would not have quantified trial quality using a scale but would have described it in full, where applicable.
Usually there is significant heterogeneity between clinical trials examining the same intervention. Had we found any studies, we would have grouped participants by clinical categories or disease stage according to the World Health Organization (WHO) (WHO 2006), as shown in Table 6 or Centers for Disease Control and Prevention (CDC) 1994 revised HIV paediatric classification system criteria (CDC 1994), as seen in Table 2. Hospital-based in-patient studies would have been considered separately from studies performed in ambulatory patients. We would have grouped people on highly active antiretroviral therapy (HAART) differently from those yet to commence HAART. We would have utilized a Chi-square test for homogeneity to test between-trial heterogeneity and the I-squared test to quantify the degree of heterogeneity. If we had found significant heterogeneity, the results would have been considered non-comparable, and we would not have performed a meta-analysis. If subgrouped studies were relatively homogeneous and it was possible to combine trial results in a meta-analysis, a random effects method would have been used. For binary outcomes, we would have calculated treatment effect as relative risks and calculated the 95% confidence intervals. For continuous outcomes, treatment effects would have been calculated as weighted mean differences, while for survival analysis, they would have been combined using the generic inverse variance. We also would have made estimates of hazard ratios where this data was provided in useable format.
Description of studies
No studies were identified that fulfilled the selection criteria
Risk of bias in included studies
No studies were identified that fulfilled the selection criteria.
Effects of interventions
This review has no analyses
In this review, no RCTs on the effectiveness of iron supplementation in reducing morbidity and mortality in HIV-infected children were found.
Although there have been several trials of iron supplementation in HIV-uninfected children, a systematic review with Cochrane-style methodology also is lacking. There was one non-Cochrane review of oral iron supplementation in non-HIV-infected children aged to 59 months who live in developing countries to ascertain the associated health benefits and risks (Iannotti 2006). In this review of 26 RCTs, varied primary outcomes were considered and there was sufficient evidence for benefits with respect to cognitive function, motor development, recovery from anaemia, and prevention of anaemia. Presupplementation iron stores and anaemia appeared to act as effect modifiers. The evidence providing support for the role of baseline iron indices and anaemia were that clear benefits followed supplementation of iron-deficient and anaemic children and the reports of adverse events related to malaria-related morbidity and poor weight gain in iron-replete children. Current practice of iron supplementation in HIV-infected adults and children is based on observational studies, especially those reporting the baseline haematological status and indices of iron storage in HIV-infected persons, including children (Eley 2002; Totin 2002; Adetifa 2006); those that report no evidence for deleterious effects of iron ( Semba 2001;Kupka 2007; Semba 2007); and others that observed development of opportunistic infection and the progression of HIV disease (Weinberg 1996; Sahli 1998; Gordeuk 2006).
There is an urgent need to conduct RCTs of iron supplementation in HIV-infected children, especially those resident in malarious areas and other parts of the world with significant overlap between prevalence of iron- and non-iron-related anaemia to provide high-quality evidence for the effectiveness of this intervention. It is key for studies to be done in settings with high burdens of paediatric HIV/AIDS and iron deficiency and iron deficiency anaemia. Ideally, these trials should include children of ages from infancy to adolescence and children already on HAART. It also will be important for these studies to conduct baseline and serial assessments of iron storage indices and haemoglobin, as these are likely effect modifiers. The outcomes of these trials must be adjusted for clinical and immunologic classification of the study population at diagnosis of HIV and also for viral loads. At the same time, it will be important to obtain evidence from unbiased reviews of trials of iron supplementation in HIV-uninfected children in malarious and non-malarious areas to obtain some comparative data for a normal population of children. Completion of a number of Cochrane reviews in progress (Martins 2001; Fraser 2006; Xiaoxi 2007; Ojukwu 2007) will assist with gathering this evidence. Apart from ensuring that evidence is collected from varied geographic settings, it is important that trials are adequately powered to examine results related to clinical outcomes, including growth, mental and physical development, haematologic profiles, susceptibility to infections, and ease of supplementation, as well as virologic and immunologic outcomes.
It is important to highlight key ethical and logistic issues and questions that may arise if these trials are to be conducted properly. It may be argued that HIV-infected children with clinical or laboratory features, or both, of iron deficiency or anaemia should be promptly treated, in view of the benefits of iron supplementation to overall mental and physical growth, and therefore not randomised to a placebo or non-iron supplementation arm of any trial. However, until there is incontrovertible evidence for or against a deleterious effect of iron on viral and immunologic outcomes, especially in this era of HAART, it will be unclear if iron supplementation falls within the ambits of acts of maleficence or beneficence in HIV-infected children and adults. Reports of resolution or improvement of anaemia in HIV-infected persons on HAART without supplementation further complicates issues of trial design and interpretation (Berhane 2004). It is common knowledge that HIV infection is associated with a chronic inflammatory response attributed both to the virus and to opportunistic infections that may be present. Serum ferritin and plasma transferrin receptor concentration (common parameters for assessing body iron status) are affected by the acute-phase response (Semba 2000; Weinberg 2001). Therefore, baseline and serial measurements of other acute phase reactants, such as C-reactive protein and alpha1-antichymotrypsin, must be performed, and where these suggest an ongoing acute-phase response, outcome data may have to be adjusted for the acute-phase reaction or even excluded. Obviously, this may lead to biased results or an inadequately powered study if these possibilities are not considered during the design of clinical trials.
Investigating the effectiveness of iron supplementation in HIV-infected children and adults is of particular relevance to developing world settings where the burden of HIV is highest because supplementation is a relatively simple intervention that readily can be translated into practice. This also applies to studies analysing supplementation with other micronutrients.
Implications for practice
The current clinical practice of iron supplementation in HIV-infected children is based on weak evidence comprising observational studies and expert opinions.
Implications for research
High-quality RCTs of iron supplementation are required in children who live in areas with significant overlap of high prevalence of HIV, malaria, and IDA. Policy makers should give priority to and allocate funds for such studies. Clinical and community settings with significant paediatric HIV cohorts are most appropriate settings for these studies.
We acknowledge the assistance of the South African Cochrane Centre and Cochrane HIV/AIDS Review Group Mentoring Programme.
Nandi Siegfried provided mentorship to Ifedayo Adetifa. We are indebted to Joy Oliver for her patience and friendliness in providing support and dealing with all our queries as well as for our literature search.
We thank Elizabeth Pienaar for her help in putting together our search strategy and with the actual search.
Data and analyses
This review has no analyses.
Last assessed as up-to-date: 29 September 2008.
Protocol first published: Issue 3, 2007
Review first published: Issue 1, 2009
Declarations of interest
Sources of support
- MRC Laboratories, Fajara, P.O. Box 273, Banjul, Gambia.
- European and Developing Countries Clinical Trials Partnership, Netherlands.
- South African Cochrane Centre HIV/AIDS Mentoring Programme, South Africa.
Medical Subject Headings (MeSH)
MeSH check words