A9.1 Is a loading dose of quinine (20 mg/kg) superior to no loading dose?
One systematic review, and one subsequent randomized controlled trial (RCT) in children found no significant difference in mortality between quinine regimens with a high initial quinine dose and those with no loading dose. However, parasite and fever clearance times were reduced in the former.
One systematic review (search date 2002, three RCTs, 92 people) (1). One subsequent RCT (2).
The systematic review found no significant difference in mortality between a group receiving a high initial dose of quinine (20 mg salt/kg or 16 mg base/kg given by the intramuscular (i.m.) route or by i.v. infusion) followed by a standard dose of quinine, and one receiving the standard dose but no loading dose (two RCTs; 2/35 (5.7%) died in the group receiving a high initial dose, 5/37 (13.5%) with no loading dose; RR: 0.43; 95% CI: 0.09-2.15) (1). One of the RCTs (39 children) found no significant difference between the two groups in mean time to recover consciousness (14 h with a high initial dose, 13 h with no loading dose, weighted mean difference (WMD) 1.0 h; 95% CI: -8.8 h to +10.8 h) (3). Parasite clearance and fever clearance times were shorter for the high initial dose group than for the group with no loading dose (parasite clearance time: two RCTs, 67 people, WMD 7.4 h; 95% CI: -13.2 h to -1.6 h; fever clearance time, two RCTs, 68 people, WMD -11.1 h; 95% CI: -20.0 h to -2.2 h). The subsequent RCT (72 children aged 8 months to15 years in Togo, 1999-2000) found no significant difference between the group receiving a high initial dose of i.v. quinine (20 mg salt/kg over 4 h, then 10 mg salt/kg every 12 h) and that receiving no loading dose (15 mg salt/kg every 12 h) in mortality (2/35 (6%) with a high initial dose, 2/37 (5%) with no loading dose, RR: 1.06; 95% CI: 0.16-7.1) (2). It also found no significant difference between the two groups for recovery of consciousness (35.5 h with a high initial dose, 28.6 h with no loading dose, WMD: +6.9 h; 95% CI: -0.6 h to +14.4 h) or time to 100% parasite clearance (48 h compared to 60 h).
The systematic review found no significant difference between the groups receiving a high initial dose of quinine and no loading dose in rate of hypoglycaemia (two RCTs; 4/35 (11%) hypoglycaemia with a high initial dose, 3/37 (8%) with no loading dose, RR: 1.39; 95% CI: 0.32-6.00) (1). In one RCT (33 people) included in the review, transient partial hearing loss was significantly increased in the group receiving a high initial dose (10/17 (59%) compared to 3/16 (19%), RR: 3.14; 95% CI: 1.05-9.38) (4). In another (39 children), there was no significant difference between the two groups in neurological sequelae (1/18 (6%) with a high initial dose, 2/21 (10%) with no loading dose, RR: 0.58; 95% CI: 0.06-5.91) (3).
A9.2 Is intramuscular quinine as effective as intravenous quinine?
One RCT in children found no significant difference between i.m. and i.v. quinine in recovery times or death. However, the study may have lacked the power to detect clinically important differences between treatments.
No systematic review. One RCT (59 children aged <12 years, Kenya, 1989-1990), which compared i.m. quinine (20 mg salt/kg loading immediately followed by 10 mg salt/kg every 12 h) with standard-dose quinine given by i.v. infusion (10 mg salt/kg every 12 h) in severe falciparum malaria (3). The trial found no significant difference in mortality, mean parasite clearance time or recovery time to drinking or walking, but may have lacked the power to detect a clinically important difference (mortality: 3/20 (15%) deaths with i.m. quinine, 1/18 (5.6%) with i.v. quinine, RR: 2.7; 95% CI: 0.3-23.7; mean parasite clearance time: 57 h compared to 58 h, WMD: -1.0 h; 95% CI: -12.2 h to +10.2 h; mean recovery times to drinking 47 h compared to 32 h, WMD: +15 h; 95% CI: -5.6 h to + 35.6 h; mean recovery times to walking: 98 h compared to 96 h, WMD: +2.0 h; 95% CI: -24.5 h to +28.5 h).
Neurological sequelae were reported in two children in the i.m. group, and one child in the i.v. group had transient neurological sequelae that were not specified (2/20 (10%) compared to 1/18 (5.6%), RR: 1.8; 95% CI: 0.2-18.2) (3).
A9.3 Is intrarectal quinine as effective as intravenous or intramuscular quinine?
One systematic review of 8 trials detected no difference in effect on parasites, clinical illness between the rectal group and either i.m or i.v. groups. Some studies however excluded patients with severe malaria.
One systematic review (search date 2005, eight RCTs, 1247 people) (5). Five trials compared intrarectal quinine with i.v quinine infusion, while 6 compared with intramuscular quinine. The systematic review found no significant difference between intrarectal with i.v. or i.m. routes for death, parasite clearance by 48 hours and 7 days, parasite clearance time, fever clearance time, coma recovery time, duration of hospitalization, and time to drinking. However trials reporting these outcomes were small, which resulted in large confidence intervals for all the outcomes.
No reported rectal irritation. Muciod stools, however, were reported in 4 patients receiving rectal quinine. No statistically significant difference between painful swelling at the site of application or pain at the site of application after administration (5).
A9.4 Is intravenous artesunate superior to intravenous quinine?
One large multi-centre trial from South-East Asia enrolling 1461 patients (including 202 children less than 15 years old) found a significant advantage of artesunate over quinine in mortality (15% vs 22%). There was an absolute reduction in mortality of 34.7% (95% CI: 18.5-47.6%; P = 0.002) in the artesunate group. Quinine was associated with hypoglycaemia (RR: 3.2, P = 0.009).
No systematic reviews: two RCTs. The first (113 adults with severe malaria, Thailand) comparing i.v. artesunate (2.4 mg/kg initially, 1.2 mg/kg 12 h later, then 1.2 mg/kg daily) to i.v. quinine (20 mg/kg initially, then 10 mg/kg every 8 h) (6). It found no significant difference between the treatments in mortality after 300 h (7/59 (12%) artesunate, 12/54 (22%) quinine, RR: 0.53; 95% CI: 0.23-1.26). It found that artesunate significantly improved parasite clearance time, but that there was no significant difference in fever clearance time or coma recovery time (parasite clearance time: 63 h with artesunate, 76 h with quinine, p = 0.019; fever clearance time: 41 h compared to 65 h, p = 0.2; coma recovery time: 17 h compared to 18 h, p = 0.6).
The second RCT is a large multi-centre trial (SEAQUAMAT study group: Bangladesh, India, Indonesia and Myanmar) with 1431 patients enrolled (7). It found that mortality of 15% (107/730 in the artesunate group was significantly lower than the 22% (164/731) in the quinine group. An absolute reduction of 34.7% (95% CI: 18.5-47.6%; P = 0.0002) in the artesunate group. There are, however, still insufficient data for children, particularly from high transmission settings.
The first RCT found that artesunate significantly reduced hypoglycaemia compared with quinine (6/59 (10%) compared to 15/54 (28%), RR: 0.37; 95% CI: 0.15-0.88) (6). One person treated with artesunate developed an urticarial rash. A similar finding was obtained in the second RCT where treatment with artesunate was well tolerated, whereas quinine was associated with hypo-glycaemia (RR: 3.2; 95% CI: 1.3-7.8; P = 0.009) (7).
A9.5 Is intramuscular artemether as effective as intravenous quinine?
Two systematic reviews and three subsequent RCTs found no significant difference in death rates between the groups receiving artemether and quinine for severe malaria.
Two systematic reviews (8, 9) and three subsequent RCTs (10-12). The first review (search date not reported, seven RCTs, 1919 adults and children) analysed individual participant data (8). It found no significant difference in mortality between i.m. artemether and quinine given by i.v. infusion or i.m injection (the latter in one RCT only) in severe falciparum malaria (mortality 136/961 (14%) with artemether, 164/958 (17%) with quinine; OR: 0.80; 95% CI: 0.62-1.02). Parasite clearance was faster with artemether than with quinine (OR: 0.62; 95% CI: 0.56-0.69). The review found no significant difference in the speed of coma recovery, fever clearance time or neurological sequelae between artemether and quinine (coma recovery time, OR: 1.09; 95% CI: 0.97-1.22; fever clearance time, OR: 1.01; 95% CI: 0.90-1.15; neurological sequelae, OR: 0.82; 95% CI: 0.59-1.15).
The second review (search date 1999, 11 RCTs, 2142 people) found a small significant reduction in mortality for i.m. artemether compared with i.v. quinine (OR: 0.72; 95% CI: 0.57-0.91) (9). However, more rigorous analysis excluding three poorer quality trials found no significant difference in mortality (OR: 0.79; 95% CI: 0.59-1.05). The review found no significant difference in neurological sequelae at recovery between the artemether and quinine groups (OR: 0.8; 95% CI: 0.52-1.25).
The first subsequent RCT (105 people aged 15-40 years with cerebral malaria in Bangladesh) compared i.m. artemether (160 mg initially, then 80 mg/kg once daily) with i.v. quinine (loading dose 20 mg/kg, then 10 mg/kg every 8 h) (10). It found no significant difference in death rates between the artemether and quinine groups (9/51 (18%) compared to 10/54 (19%), OR: 0.94; 95% CI: 0.35-2.55). Mean fever clearance time and coma recovery time were significantly longer for artemether than for quinine (fever clearance time: 58 h compared to 47 h, WMD: 11.0 h; 95% CI: 1.6-20.4 h; coma recovery time: 74 h compared to 53 h, WMD: 20.8 h; 95% CI: 3.6-38.0 h). There was no significant difference in mean parasite clearance time between artemether and quinine (52 h compared to 61 h, WMD: 8.6 h; 95% CI: 22.5 h to +5.3 h).
The second subsequent RCT (41 children with severe malaria in Sudan, 40 analysed) compared i.m. artemether (3.2 mg/kg loading dose, then 1.6 mg/kg once a day) with i.v. quinine (loading dose 20 mg/kg, then 10 mg/kg every 8 h) (11). It found that artemether significantly increased fever clearance time but found no significant difference between artemether and quinine in time to parasite clearance (mean fever clearance time: 30.5 h with artemether, 18 h with quinine, p = 0.02; mean parasite clearance time: 16 h compared to 22.4 h, p > 0.05). There were no deaths in the artemether group, one child died with quinine (0/20 (0%) compared to 1/21 (5%), p value not reported).
The third subsequent RCT (77 comatose children aged 3 months-15 years with cerebral malaria) compared i.m. artemether (1.6 mg/kg every 12 h) with i.v. quinine (10 mg/kg every 8 h) (12). It found no significant difference in death rates between the artemether and quinine groups (3/38 (8%) compared to 2/39 (5%), p value not reported). There was no significant difference between the two groups in mean fever clearance time, coma recovery time and parasite clearance time (fever clearance time 31 h compared to 36 h; coma recovery time: 21 h compared to 26 h; parasite clearance time 36 h compared to 41 h; p values not reported for any comparison).
The two systematic reviews (8, 9) and one of the subsequent RCTs (3) found no significant difference in neurological sequelae between the artemether and quinine groups (systematic reviews, see the Benefits section; subsequent trial 3/51 (6%) with artemether, 1/54 (2%) with quinine, RR: 3.18; 95% CI: 0.34-29.56). However, in the first review, rates for the combined outcome of death or neurological sequelae were lower for artemether than for quinine (OR: 0.77; 95% CI: 0.62-0.96, p = 0.02) (8).
The second subsequent RCT found that one child treated with quinine developed hypoglycaemia (0/20 (0%) with artemether, 1/21 (5%) with quinine, (11). It reported no neurological problems in either treatment group after 28 days of follow-up.
The third subsequent RCT found no significant difference in transient neurological sequelae between the artemether and quinine groups (2/38 (5%) compared to 1/39 (3%), (12).
The third subsequent randomized controlled trial did not use loading doses of either artemether or quinine at the beginning of treatment (12). There was a fourth subsequent RCT (52 people) (13). However, it was not clear whether participants had severe malaria, and outcomes were poorly reported.
A9.6 Is intramuscular artemotil as effective as intravenous quinine?
Cochrane Review (search date August 2004). Two small trials (n = 194) met the inclusion criteria (14). Both trials compared i.m. artemotil with quinine given by i.v. infusion in children with cerebral malaria and reported on similar outcomes. There was no statistically significant difference in the number of deaths (RR: 0.75; 95% CI: 0.43-1.30, n = 194, 2 trials), neurological complications (RR: 1.18;, 95% CI: 0.31-4.46, n = 58, 1 trial), or other outcomes including time to regain consciousness, parasite clearance time and fever clearance time. The meta-analyses lack the statistical power to detect important differences.
A9.7 Is rectal artemisinin as effective as intravenous quinine?
One systematic review of small RCTs found no significant difference in mortality between rectal artemisinin and quinine given by i.v. infusion.
One systematic review (search date 1999, three RCTs) comparing rectal artemisinin with i.v. quinine in severe malaria (9). Two RCTs were conducted in 1996-1997. Meta-analysis showed lower mortality with artemisinin and quicker coma recovery time, but the difference was not significant (mortality, three RCTs, 9/87 (10%) with artemisinin, 16/98 (16%) with quinine, RR: 0.73; 95% CI: 0.35-1.50; coma recovery, two RCTs, 59 people, WMD: -9.0 h; 95% CI: -19.7 h to +1.7 h). Fever clearance times were not significantly different (no figures provided).
One RCT in children found that artemisinin significantly reduced the risk of hypoglycaemia compared with quinine (3/30 (10%) with artemisinin, 19/30 (63%) with quinine, RR: 0.16; 95% CI: 0.05-0.48) (15).
A9.8 Pre-referral treatment with rectal artesunate: should rectal artesunate be used in preference to quinine?
There are no data from trials with sufficient statistical power to assess differences in mortality following treatment with rectal artesunate and quinine in people with moderate or severe malaria. The objective of the trials that have been conducted was to establish the safety and efficacy of rectal artesunate as pre-referral treatment where there is no access to parenteral treatment. Comparisons between rectal artesunate and i.v. artesunate or i.v., i.m. quinine have been carried out to assess parasitological and clinical response in the 12 or 24 hours immediately after treatment (16, 17).
Two randomized, open-label Phase II and three randomised open label Phase III studies have been conducted in people with moderately severe malaria, i.e. patients who could not take drugs by mouth but did not have features of severe malaria and its complications (16, 17). Patients in the artesunate group in the Phase III studies were rescued if their parasitaemia did not decline to below 60% of baseline parasitaemia or if they deteriorated clinically and developed features of severe malaria, convulsions or coma within 24 hours of treatment.
Artesunate had a superior effect in all efficacy measures immediately after treatment. In children treated with artesunate, 80/87 (92%) had a parasite density lower than 60% of baseline, compared to 3/22 (14%) of those who received quinine (RR: 0.09; 95% CI: 0.04-0.19, p <0.0001). In adult, parasitaemia at 12 hours was lower than 60% of baseline in 26/27 (96%) in the artesunate group, compared to 3/8 (38%) in the quinine group. (RR: 0.06; 95% CI: 0.01-0.44, p <0.001). The differences were more significant at 24 h. Artesunate and/or dihydroartemisinin were detected in plasma within 12 h in all adults and in 84/87 of the children.
A single administration of artesunate suppositories at a dose of 10 mg/kg was well tolerated in both children and adults. There was no significant difference in frequency of adverse events (defined as any new symptom, worsening of any existing symptom, sign or abnormal laboratory value) between treatment groups. Other than local reactions at the site of the i.m. quinine injection in three adult patients, the few adverse events that occurred could have been attributable to falciparum malaria or to pre-existing disease.
A9.9 Should dexamethasone be given routinely?
One systematic review found no significant difference in mortality between dexamethasone and placebo, but gastrointestinal bleeding and seizures were more common with dexamethasone.
One systematic review (search date 1999, two RCTs, 143 people with severe cerebral malaria treated with quinine), which compared dexamethasone with placebo over 48 h (18). One RCT was conducted in Indonesia, the other in Thailand. The review found no significant difference in mortality (14/71 (20%) with dexamethasone, 16/72 (25%) with placebo, RR: 0.89; 95% CI: 0.47-1.68). One RCT found a longer mean time between start of treatment and coma resolution with dexamethasone (76.0 h compared to 57.0 h, p <0.02) (18), but the other found no significant difference (83.4 h compared to 80.0 h, WMD: +3.4 h; 95% CI: -31.3 h to +38.1 h) (20).
The review found that dexamethasone significantly increased gastrointestinal bleeding and seizures compared with placebo (gastrointestinal bleeding 7/71 (10%) with dexamethasone, 0/72 (0%) with placebo, RR: 8.17; 95% CI: 1.05-63.6; seizures 1/71 (15.5%) compared to 3/72 (4%), RR: 3.32; 95% CI: 1.05-10.47) (18).
No effect of the steroid on mortality was shown, but the trials were small. Its effect on disability was not reported.
A9.10 Is an initial blood transfusion effective for treating malarial anaemia?
One systematic review found insufficient data to be sure whether routine administration of blood to clinically stable children (no respiratory distress or cardiac failure) with severe anaemia in endemic malarious areas reduces death, or results in higher haematocrit measured at one month. The review found no significant difference between transfusion and no transfusion for the combined outcome of death or severe adverse events. Transmission of hepatitis B or HIV was not reported. No RCTs examining the effects of transfusion in adults with malaria.
One systematic review (search date 1999, RCTs, 230 children with malarial anaemia; packed cell volume range 12-17%) (21). The first RCT (116 children, United Republic of Tanzania) compared initial blood transfusion with conservative treatment, and the second (114 children, The Gambia) compared blood transfusion with iron supplements. Both trials excluded children who were clinically unstable with respiratory distress or signs of cardiac failure. Meta-analysis found fewer deaths in the transfused children, but the difference was not significant (1/118 (1%) with transfusion, 3/112 (3%) with control, RR: 0.41; 95% CI: 0.06-2.70). No RCTs examining the effects of transfusion in adults with malaria.
The systematic review found that coma and convulsions occurred more often after transfusion (8/118 (6.8%) with transfusion, 0/112 (0%) without, RR: 8.6; 95% CI: 1.1 to 66.0) with seven of the eight adverse events occurring in one of the RCTs. Meta-analysis combining deaths and severe adverse events found no significant difference between children who received transfusions and those who did not (8/118 (7%) with transfusion, 3/112 (3%) without transfusion, RR: 2.5; 95% CI: 0.7-9.3). Transmission of hepatitis B or HIV was not reported (21).
Studies were small and loss to follow-up was greater than 10%; both of these factors are potential sources of bias. In the first RCT, one child in the transfusion group and one child in the conservative treatment group required an additional transfusion after clinical assessment. In the second RCT, 10 children allocated to receive iron supplements later required transfusion when packed cell volume fell below 12% or they showed signs of respiratory distress.
A9.11 Should phenobarbital be given to patients?
Cochrane Review, search date 2004 (22). Three RCTs with a total of 573 participants met the inclusion criteria. All three compared phenobarbital with placebo or no treatment. In the two trials with adequate allocation concealment, death was more common in the anticonvulsant group (RR: 2.0; 95% CI: 1.20-3.33, fixed-effect model). In all three trials, phenobarbital was associated with fewer convulsions than placebo or no treatment (RR: 0.30; 95% CI: 0.19-0.45, fixed-effect model).
A9.12 Hyperparasitaemia in non-immune or semi-immune populations
Since the classic paper by Field and Niven in 1937 (23), patients with high parasite counts have been known to be at increased risk of dying. Working in Peninsular Malaysia they showed that P. falciparum parasite counts of >100 000/µl of blood (2%) were associated with an increased risk of mortality, while patients with counts of >500 000/µl had a more than 50% chance of dying. Following this observation, hyperparasitaemic patients have been considered to be at high risk. They are often classified as having severe malaria and accordingly treated with parenteral antimalarial drugs wherever possible.
Many hyperparasitaemic patients have evidence of vital organ dysfunction but there is a subgroup of individuals in whom no other manifestations of severe disease occur. These patients have symptoms and signs compatible with a diagnosis of uncomplicated malaria in association with a high parasite count.
They usually have a predominance of young ring forms in the peripheral blood smear, which suggests that the sequestered biomass is relatively small compared with the circulating parasite numbers. On the Thai-Burmese border, the mortality of falciparum malaria in patients with counts of parasitized red blood cells <4% was approximately 0.1%, whereas with counts >4% but without vital organ dysfunction, the mortality was 3% (24). In another study in eastern Thailand, 48 hyperparasitaemic patients were inadvertently included in a cohort of 224 patients (mainly adult men) being followed up after receiving mefloquine treatment for uncomplicated disease in order to identify predictors of treatment failure. Young age, low admission haemoglobin, a history of diarrhoea at the start of treatment or soon after, three or more mefloquine treatments within the preceding year and hyperparasitaemia on an admission blood film were all predictors of subsequent treatment failure (25). The study took place at a time when mefloquine resistance was becoming well established. The authors concluded that, in areas of multidrug resistance, the initial parasite load becomes a critical determinant of treatment failure because strains have intermediate susceptibility to drugs. Patients with little or no immunity to malaria (i.e. children) are at increased risk of treatment failure. In a review of the predictors of treatment failure in a large series of patients studied on the western border of Thailand, the relative risk of treatment failure was increased in patients with parasite counts >10 000/µl (RR: 1.36; 95% CI: 1.12-1.66) (25). In Uganda, it was found that pretreatment parasitaemia >100 000/µl was associated with treatment failure in adults (OR for failure: 2.26; 95% CI: 1.30-3.92), but the authors stated that this was due to a strong association with body temperature in a multivariate analysis (26).
There is no uniformly agreed definition of hyperparasitaemia. Some publications use a parasite count >100 000/µl of blood. Others use a proportion of parasitized red blood cells >4%, >5% or >10%. In areas of high malaria transmission, mortality rates in patients with 4% or 5% parasitaemia are considerably lower than in areas of low transmission because of the influence of host immunity.
Is parenteral treatment better?
There are very few studies specifically looking at treatment failure as an outcome in hyperparasitaemic patients. A study working with hyperparasitaemic (>4%) patients in a low-transmission setting, compared oral (artesunate for 3 days + mefloquine on day 2) with i.v. therapy (loading-dose quinine regimen for 24 h followed by mefloquine). Patients in the oral therapy group had significantly faster parasite and fever clearance and shorter hospital stays than those in the intravenous therapy group. No patients progressed to severe disease. However, 28-day cure rates for the oral therapy group were poorer than in non-hyperparasitaemic controls receiving the same treatment (70% compared to 96%). It was concluded that patients with hyperparasitaemia in areas of multidrug resistance need a longer course of treatment to deal with the larger parasite biomass (27).
In one study from a high-transmission setting, 84 hyperparasitaemic (>5%) children aged 1-10 years were assigned at random to receive either 5 days of i.m. artemether or a single oral dose of mefloquine. Cure rates assessed at day 14 were similar in the two groups (98% at day 14 for artemether, 100% at day 28 for mefloquine), although parasite clearance was faster in the artemether group. Four patients in the mefloquine group were excluded because of vomiting, but no patients progressed to severe disease. The authors concluded that oral therapy may be given safely to hyperparasitaemic patients in Africa provided that they can tolerate it orally and the parasites are fully sensitive to the drug (28).
Another study from an area of stable transmission showed rapid parasite clearance in all 32 patients with parasitaemias of 5-35% (mean 10.8%) following oral amodiaquine therapy. Follow-up was to day 14 only and there was no comparative treatment arm. The authors concluded that hyperparasitaemic patients can safely be treated with oral therapy if they have no features of severe disease and can be closely monitored in the first 3 days (29). The same conclusion was reached from another cohort study from an area of stable malaria transmission (30).
Is duration of treatment important?
A study of 100 hyperparasitaemic (>4%) adults and children on the Thai-Burmese border compared treatment with oral artesunate given alone for 5 days with artesunate given for 5 days + mefloquine on day 2. Recrudescence (i.e. treatment failure) rates were 36% and 6% respectively. Two further hyperparasitaemic treatment groups were then investigated: 34 patients with recrudescent malaria following treatment with artesunate + mefloquine who then received artesunate for 7 days; and 132 patients from a malaria vaccine study with primary infections who received artesunate for 7 days + mefloquine on day 2. In these two groups, which were non-randomized and not strictly comparable, recrudescence rates were 26% and 7%, respectively. The authors concluded that duration of treatment is important, particularly when using artemisinin derivatives with short half-lives; that the addition of mefloquine improved cure rates; and that it is important that the parasite counts are reduced massively by artesunate before using mefloquine so that selection pressure for resistance to mefloquine can be minimized (31).
Patients with extreme hyperparasitaemia (>20%)
There have been no treatment studies looking at these patients specifically. They are often treated as having severe malaria even if they can take oral treatment. In low-transmission settings this group of patients was excluded from hyperparasitaemia studies because of the risks of clinical deterioration. Current practice for patients with parasitaemia >20% in these settings is to give parenteral artesunate (or parenteral quinine if artesunate is not available), but this approach is not based on evidence (E. Ashley, personal communication). In Nigeria, where there are higher transmission rates, an upper limit to hyperparasitaemia was not used but the patients with extreme hyperpara-sitaemia were not analysed as a separate group, so no data specific to this group are available.
Conclusions and recommendations
Hyperparasitaemia carries an increased risk of mortality in falciparum malaria. The activity of artemisinin derivatives against circulating parasites makes them particularly effective in hyperparasitaemic patients who have no other signs of severity. Treatment failure rates are higher in hyperparasitaemic patients. This is also a potential source of anew drug resistance.
Hyperparasitaemic patients with no signs of severe disease may be treated with oral artemisinin derivatives provided that the following apply:
• Patients must be monitored closely for the first 48 h after initiation of treatment.
• They must tolerate the drug(s) well, i.e. without diarrhoea or vomiting.
• For regimens containing mefloquine, the mefloquine should be given on day 2 when it is better tolerated, with a lower incidence of early vomiting, rather than on day 0.
• If possible, additional oral artemisinin derivative should be given so that the total treatment course is 5-7 days. This has been studied only for arte-sunate + mefloquine (artesunate 4 mg/kg immediately then 2 mg/kg/day for a further 4-6 days + mefloquine 25 mg/kg given as a split dose after the second day). Alternatively, the first dose of artemisinin derivative can be given parenterally or rectally to ensure adequate absorption.
Non-immune patients with parasitaemia >20% should continue to receive parenteral therapy wherever possible, as there is no evidence for or against using oral treatment in this group and the risks are high.
1. Lesi A, Meremikwu M. High first dose quinine for treating severe malaria. In: The Cochrane Library, Issue 2, 2003. (Oxford: Update Software. Search date 2002; primary sources Cochrane Infectious Diseases Group specialised trials register, Cochrane Controlled Trials Register, Medline, Embase, Lilacs, conference proceedings, researchers working in the field, and hand searches of references).
2. Assimadi JK et al. [Intravenous quinine treatment of cerebral malaria in African children: comparison of a loading dose regimen to a regimen without loading dose]. Archives de Pediatrie, 2002, 9:587-594 (in French).
3. Pasvol G et al. Quinine treatment of severe falciparum malaria in African children: a randomized comparison of three regimens. American Journal of Tropical Medicine and Hygiene, 1991, 45:702-713.
4. Tombe M, Bhatt KM, Obel AOK. Quinine loading dose in severe falciparum malaria at Kenyatta National Hospital, Kenya. East African Medical Journal, 1992, 69:670-674.
5. Eisenhut M, Omari AA. Intrarectal quinine for treating Plasmodium falciparum malaria. Cochrane Database Syst Rev, 2005, 25(1): CD004009
6. Newton PN et al. Randomized comparison of artesunate and quinine in the treatment of severe falciparum malaria. Clinical Infectious Diseases, 2003, 37:7-16.
7. South-East Asian Quinine Artesunate Malaria Trial (SEAQUAMAT) group. Artesunate versus quinine for treatment of severe falciparum malaria: a randomized Trial. Lancet, 2005, 366:717-725.
8. Artemether Quinine Meta-analysis Study Group. A meta-analysis using individual patient data of trials comparing artemether with quinine in the treatment of severe falciparum malaria. Transactions of the Royal Society Tropical Medicine and Hygiene, 2001, 95:637-650. (Search date not reported; primary sources Medline, Cochrane, and discussions with an international panel of malaria clinical investigators).
9. McIntosh HM, Olliaro P. Artemisinin derivatives for treating severe malaria (Cochrane Review). In: The Cochrane Library, Issue 2, 2001. (Oxford: Update Software. Search date 1999; primary sources Cochrane Infectious Diseases Group Trials Register, Medline, Bids, Science Citation Index, Embase, African Index Medicus, Lilacs, hand searches of reference lists and conference abstracts, and contact with organizations and researchers in the field and pharmaceutical companies).
10. Faiz A et al. A randomized controlled trial comparing artemether and quinine in the treatment of cerebral malaria in Bangladesh. Indian Journal of Malariology, 2001, 38:9-18.
11. Adam I et al. Comparison of intramuscular artemether and intravenous quinine in the treatment of Sudanese children with severe falciparum malaria. East African Medical Journal, 2002, 79:621-625.
12. Satti GM, Elhassan SH, Ibrahim SA. The efficacy of artemether versus quinine in the treatment of cerebral malaria. Journal of the Egyptian Society of Parasitology, 2002, 32:611-623.
13. Singh NB et al. Artemether vs quinine therapy in Plasmodium falciparum malaria in Manipur - a preliminary report. Journal of Communicable Diseases, 2001, 33:83-87.
14. Afolabi BB, Okoromah CN. Intramuscular arteether for treating severe malaria. In: The Cochrane Database of Systematic Reviews 2004, Issue 4.
15. Birku Y, Makonnen E, Bjorkman A. Comparison of rectal artemisinin with intravenous quinine in the treatment of severe malaria in Ethiopia. East African Medical Journal, 1999, 76:154-159.
16. Krishna S et al. Bioavailability and preliminary clinical efficacy of intrarectal artesunate in Ghanaian children with moderate malaria. Antimicrobial Agents and Chemotherapy, 2001, 45:509-516.
17. Barnes KI et al. Efficacy of rectal artesunate compared with parenteral quinine in initial treatment of moderately severe malaria in African children and adults: a randomised study. Lancet, 2004, 363:1598-1605.
18. Prasad K, Garner P. Steroids for treating cerebral malaria. In: The Cochrane Library, Issue 2, 2003. (Oxford: Update Software. Search date 1999; primary sources Trials Register of the Cochrane Infectious Diseases Group and Cochrane Controlled Trials Register.)
19. Warrell DA et al. Dexamethasone proves deleterious in cerebral malaria. A double-blind trial in 100 comatose patients. New England Journal of Medicine, 1982, 306:313-319.
20. Hoffman SL et al. High-dose dexamethasone in quinine-treated patients with cerebral malaria: a double-blind, placebo-controlled trial. Journal of Infectious Diseases, 1988, 158:325-331.
21. Meremikwu M, Smith HJ. Blood transfusion for treating malarial anaemia (Cochrane Review). In: The Cochrane Library, Issue 3, 2001. (Oxford: Update Software. Search date 1999; primary sources Trials Register of the Cochrane Infectious Diseases Group, Embase, African Index Medicus, Lilacs, hand searches of reference lists, and contact with experts.)
22. Meremikwu M, Marson AG. Routine anticonvulsants for treating cerebral malaria. The Cochrane Database of Systematic Reviews 2002, Issue 2. (Article No. CD002152. DOI: 10.1002/14651858.CD002152).
23. Field JW, Niven JC. A note on prognosis in relation to parasite counts in acute subtertian malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1937, 30:569-574.
24. Fontanet AL, Walker AM. Predictors of treatment failure in multiple drug-resistant falciparum malaria: results from a 42-day follow-up of 224 patients in eastern Thailand. American Journal of Tropical Medicine and Hygiene, 1993, 49:465-472.
25. ter Kuile FO et al. Predictors of mefloquine treatment failure: a prospective study of 1590 patients with uncomplicated falciparum malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1995, 89:660-664.
26. Dorsey G et al. Predictors of chloroquine treatment failure in children and adults with falciparum malaria in Kampala, Uganda. American Journal of Tropical Medicine and Hygiene, 2000, 62:686-692.
27. Luxemburger C et al. Oral artesunate in the treatment of uncomplicated hyperparasitemic falciparum malaria. American Journal of Tropical Medicine and Hygiene, 1995, 53:522-525.
28. Sowunmi A, Walker O, Akindele JA. Severe falciparum malaria: survival with oral therapy. Tropical and Geographical Medicine, 1992, 44:352-354.
29. Ndounga M, Basco LK. Rapid clearance of Plasmodium falciparum hyper-parasitaemia after oral amodiaquine treatment in patients with uncomplicated malaria. Acta Tropica, 2003, 88:27-32.
30. Sowunmi A et al. Open comparison of artemether and mefloquine in uncomplicated Plasmodium falciparum hyperparasitaemia in children. Annals of Tropical Paediatrics, 1996, 16:5-9.
31. Price R et al. Artesunate and mefloquine in the treatment of uncomplicated multidrug-resistant hyperparasitaemic falciparum malaria. Transactions of the Royal Society of Tropical Medicine and Hygiene, 1998, 92:207-211.