Saturday, July 25, 2015

Randomised trials in child health

By Esther Nakkazi

The latest research on child health in developing countries with evidence derived from all the randomized trials published over the last year 2014-15 are in a booklet available at

This year there were 245 publications from randomised or controlled trials, more than in any previous year. These came from all regions of the world, mostly from developing country researchers. The summarise are aimed at making this information widely available to paediatricians, child health nurses, midwives, researchers, students and administrators in places where up-to-date health information is hard to find.

"We hope it will be helpful in reviewing treatment guidelines and clinical and public health approaches, and in teaching about paediatrics and evidence-based medicine," said Prof Trevor Duke, from the Centre for International Child Health, University of Melbourne and MCRI Intensive Care Unit, Royal Children’s Hospital Melbourne, Victoria, Australia. 

Prof Duke says they hope  that such information will be helpful in reviewing treatment policies, clinical practice and public health strategies. The booklet is compiled annually, using the PubMed search engine. The search strategy is reproducible by anyone with access to the Internet, through

The method of the Randomized Trial is the Gold Standard for determining attributable benefit or harm from clinical and public health interventions. So Randomized controlled trials (RCTs) when done properly eliminate bias and confounding. 

However their results should not be accepted uncritically and they should be evaluated for quality and validity. Before the result of an RCT can be generalized to another setting there must be consideration of the wider applicability, feasibility and potential for sustainability.

Here are some of the summaries on child health in Africa found in the 2014-15 booklet;

In a high-mortality setting in Kenya where co-morbidities are common, among children with non-severe pneumonia, oral amoxicillin was non-inferior to intravenous benzylpenicillin, and failure rates at day 14 were 13.5 and 16.8% respectively. In Brazil oral amocycillin given 2 times per day was as effective as 3 times per day in treating non-severe pneumonia with treatment failure rates of 23% and 22% respectively.

Children in African hospitals with severe anaemia were more likely to die in the first 24 hours (case fatality rate 13%) than those with mild or moderate anaemia (7-8%). Children with severe anaemia who were not transfused at 2.5 hours had a much higher risk of dying than those who received blood early.

In rural children in India, Pakistan and Zambia, an early developmental intervention taught to parents over 3 years improved cognitive abilities regardless of the type of development risk the child faced.

Using a test of “intestinal permeability”, the lactulose: mannitol urinary excretion test, among children at risk of environmental enteropathy, zinc or albendazole reduced the apparent progression of intestinal permeability.

Among African children with prolonged convulsions use of intra-rectal diazepam was more effective in controlling seizures than sublingual lorazepam.

In adolescents and adults in sub-Saharan Africa with HIV and first-line treatment failure, use of a nucleoside reverse-transcriptase inhibitor was more effective as a ritonivir-boosted protease inhibitor (lopinavir-ritonavir), and as effective as combined NRTI and lopinavir-ritonavir, in achieving good HIV control (no stage 4 events, CD4>250, viral load<10,000 copies /ml at 96 weeks of observation).

In HIV exposed, uninfected infants in Kenya and South Africa, not breast-feeding was associated with a significantly increased risk of serious infectious events in the first 3 months of life.

In Zimbabwe, Nigeria, Malawi and South Africa, trials of the implementation of “Option B+”, which provides all HIV-infected pregnant and breast-feeding women with lifelong combination ART, have been planned and are underway.

In Cameroon, mobile-phone text messaging and phone call reminders increased attendance for HIV exposed or infected children.

A controlled trial of wearing shoes failed to reduce hookworm, because those in the control arm also acquired shoes! Wearing shoes in either arm was associated with a lower risk of hookworm infection.

Among children in Tanzania infected with Trichuris trichura, the use of albendazole and oxantel pamoate, or albendazole and ivermectin, were more effective than the albendazole and mebendazole, or mebendazole alone.

In Ghana, providing rapid diagnostic tests for malaria along with realistic training markedly increased the prescription of rational therapy, and in Camaroon use of RDTs reduced the costs of health care in a study which helped define the best type of health worker training.

A meta-analysis of trials of intermittent preventative therapy for malaria on the effect on anaemia showed a modest protective effect only.

Among children in Malawi treated for malaria with chloroquine-azithromycin, the incidence of subsequent respiratory and gastrointestinal infections was lower than those treated with chloroquine alone.

Among Ugandan children the use of dihydroartemisinin-piperaquine compared with artemether-lumefantrine reduced the risk of recurrent malaria and hospitalisations over the 84 days of follow-up.

Among children with sickle-cell disease, malaria parasite clearance was slower than for children without SCD when treated with artemisinin-based therapies.

In a large cluster RCT of community-based treatment of moderate malnutrition in Burkina Faso, the giving of locally produced ready-to-use supplemental feeds resulted in better weight gain than merely counselling parents about appropriate foods.

In Kenya, Mozambique and Tanzania, mothers receiving intermittent preventative therapy for malaria with mefloquine had significantly lower rates of malarial parasitaemia, placental malaria and non-obstetric hospital admissions than mothers receiving placebo, but those who received mefloquine had higher rates of perinatal mother-to-child transmission of HIV. This was an exploratory finding with potential confounding, but requires further investigation.

In Malawi, a large trial of maternal nutrient supplementation with lipid-based nutrient supplementation failed to show improved birth size or child growth in the first 18 months of life. However in another large trial in Ghana, birth weight was greater (+85g) and risk of low birth weight less with lipid-based nutrient supplementation.

In a large trial in rural Tanzania, home-based counselling of newborn care practices by volunteers improved several practices, including clean cord care and exclusive breast-feeding.

In 6 countries in South America, Asia and Africa, a trial of antenatal steroids fopr pregnant women at risk of preterm birth did not reduce mortality in those who delivered preterm, but increased neonatal and maternal sepsis and increased overall neonatal mortality.

In Democratic Republic of Congo, Kenya and Nigeria, the community based treatment of low risk but possible bacterial infection in newborns with simplified antibiotic regimens which included oral amoxicillin instead of injectable penicillin were no different in effect on newborn sepsis. Similarly for infants up to 3 months of age with fast breathing only, oral amoxicillin was as effective as injectable penicillin and gentamicin. Both trials were done in populations at very low risk of serious bacterial infection.

In a large trial in 55 villages in Burkina Faso, the implementation of an agriculture, nutrition and health behaviour program run by Helen Keller International reduced wasting, diarrhoea and anaemia.

In a large meta-analysis of 30 trials, praziquantel was the most effective drug for treating urinary schistosomiasis, however the proportion of patients cured varied from 22-83%, and trials of combination therapy with other agents is indicated. There is still no appropriate formulation of praziquantel for young children.

In a trial of shortened tuberculosis drug regimens, use of a 4-month regimen that included moxifloxacin was significantly less effective than the standard regimen 2RHZE/4RH. At this stage shortening TB treatment to less than 6 months is not of proven efficacy.

In South Africa, influenza vaccine given to pregnant HIV-positive and HIV-negative women provided partial protection (around 50% efficacy) for them, and protection for the infants the infants were HIV-unexposed. There was no protection of giving maternal influenza vaccine to infants who were HIV-infected or exposed.

In 11 African sites 3 doses of the RTS,S/AS01 malaria vaccine given to infants provided 40-50% protection against clinical malaria, 34% protection against severe malaria and 19% protection against all-cause hospitalisation.

There were several large trials of neonatal vitamin A supplementation reported on in 2014-15, finding minimal or no effect on mortality (NeoVitA trials). In one trial in India of over 40,000 newborns randomised to vitamin A 50,000 U or placebo, vitamin A showed a modest and non-significant lower mortality (-3 per 1000, 95% CI -6% to 0.1) in the first 6 months of life. In similar trials in Ghana and Tanzania involving 22,000 and 32,000 newborn infants respectively, the mortality risk was also not significantly different in the vitamin supplemented group. Bulging fontanelle was reported as an adverse effect in <1% of newborns given vitamin A. Trials from Guinea Bissau also confirmed no beneficial effect of neonatal vitamin A supplementation.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.