The study was carried out at the Sickle cell Clinic situated in Mulago, Uganda's national referral hospital in Kampala, from October 2006 to February 2007. Malaria transmission in this area is mesoendemic and perennial but peaks during the rainy seasons (from March - May, and September - November) 17. The annual inoculation rate ranges from nine to 10 infective bites per year.

Children between six months and 12 years of age attending the sickle cell clinic were consecutively selected and enrolled if they met the inclusion criteria, that is: 1) living within 10 kilometers from the Hospital; 2) having a negative malaria smear; 3) parents or guardians consented to participation. Children with a history of allergy to sulphonamides were excluded; those on co-trimoxazole prophylaxis or those with any severe illnesses needing admission. The study was approved by Makerere University Faculty of Medicine Ethics and Research Committee and Uganda National Council for Science and Technology.

This was a double-blind randomized controlled trial, where children between six months and 12 years of age attending the sickle cell clinic were randomized to receive either monthly SP or weekly chloroquine. The sample size of 242 was based on the assumption that the prevalence of malaria would 29.4% in the chloroquine group and 11.1% in the SP group 1819 with an α = 0.05, a power of 90% and an estimated 10% loss to follow.

A paediatrician not involved in the study generated a set of 242 random numbers. These numbers were then assigned to two codes in blocks of 4 to 12. Each participant was assigned a serial enrolment number and a random number corresponding to the treatment code. The pills were packed in an opaque envelope labeled with both the study serial number and a random number corresponding to the intervention. The drugs were packed in envelopes and released to the study team only after an eligible participant had been enrolled. The drugs and placebo were similar in colour and shape. They were manufactured by the same company (Kampala Pharmaceutical Industries) and labeled using two codes. Each (CQ) sugar coated tablet contained 150 mg chloroquine base and SP tablet contained sulphadoxine 500 mg and pyrimethamine 25 mg. The treatment code was concealed to the study participants, the treatment nurse and the investigators. The randomization code was only released to the investigators following completion of data analysis.

The drug were administered by a treatment nurse according to the body weight of the patients as follows: chloroquine 5 mg/kg and sulphadoxine(25 mg/kg)-pyrimethamine (1.25 mg/kg). Participants were observed for 30 minutes after drug administration and treatment re-administered if vomiting occurred.

Patients were followed up weekly for one month. At each visit, a finger prick for malaria parasites and parasite density estimation was done, in addition to haemoglobin estimation by Drabkin's method. Parents were requested to report any adverse events to the investigators. A malaria episode was considered if the study participant had fever documented by a temperature of ≥ 37.5°C and presence of any Plasmodium falciparum parasite. Children who developed malaria were treated according to the national treatment guidelines at the time (artemether-lumefantrine for uncomplicated malaria, and quinine for severe malaria).

Analysis was by intention to treat: that is all patients randomized were analysed.

The occurrence of at least one episode of malaria in the two treatment groups was compared using Odds Ratios and 95% confidence interval. The student's t-test was used for continuous variables. Logistic regression analysis was used to predict variables affecting the occurrence of malaria.

During this period 278 patients were screened and 36 patients did not fulfill the inclusion criteria. Two hundred and forty two were randomized to either SP (monthly) or chloroquine weekly (120 and 122 respectively). Seven patients were lost to follow up from the SP arm and 8 from the CQ arm. All these patients were lost follow up during the first week after enrollment. Five patients from each group that were lost to follow-up relocated to other places during the study period. The rest had different reasons such as a guardian's busy work schedule or having a sick guardian. Two hundred and twenty seven patients were followed up for one month (113 SP arm and 114 CQ arm). Baseline and laboratory characteristics (Table 1) were similar in both treatment groups implying successful randomization.

Children receiving weekly chloroquine were two times more likely to get malaria compared to those on SP. This shows that SP was more protective, reducing the prevalence of malaria by 50% compared to chloroquine. This difference in efficacy between the two drugs could be explained by the fact that resistance to therapeutic treatment of chloroquine (60%-80%) is higher compared to SP (18% - 25%) in Uganda 1220.

The other reason is that SP has a longer half life and its terminal elimination phase normally exceeds the minimum parasiticidal concentrations (lowest concentrations that give maximum effect). In contrast this does not occur for chloroquine 21.

While SP may have failed as a treatment drug it is still effective for prophylaxis 22. Several studies from Ghana, Mozambique and Tanzania have documented its usefulness in prophylaxis and particularly in reducing malaria episodes even in areas where resistance is high 1622.

In the current study, 24.6% of children receiving chloroquine contracted at least one episode of malaria compared to 40% reported in the Ethiopian study 13. This is probably due to a relatively lower resistance to chloroquine therapeutic treatment reported in Uganda than that reported from Ethiopia 2324.

In our study, SP reduced malaria related admissions by 50% compared to chloroquine, but this was not statistically significant. This suggests that SP may have a role in reducing malaria related admissions. This is in contrast to results from Ghana, where the reduction in malaria related admissions by SP was 39% in comparison to placebo among children without sickle anaemia20.

Although all cause admissions varied, the percentage of admissions was higher in the chloroquine group (9% versus 5%). Of note, admissions due to lobar pneumonia (1), painful crisis (3) only occurred in those receiving chloroquine. This might be because sulphonamides may offer an anti-bacterial effect against the diseases as has been reported among children with HIV infection 25. This implies that SP may reduce morbidity from other causes compared with chloroquine. Nonetheless, the reduction in the admissions was not statistically significant and defers from findings reported from Tanzania, and Nigeria 1526.

Drug side-effects such as vomiting, skin rash, pruritus, blurred vision were documented in a small proportion of the children (range 0.8% to 11.5%) 1520. No serious adverse effects were reported in the two treatment arms and this is consistent with findings in other studies from Tanzania, Mozambique and Ghana 1622.