- Email: [email protected]
Intermittent sulphadoxine-pyrimethamine to prevent severe anaemia secondary to malaria in pregnancy: a randomised placebo-controlled trial
ARTICLES
Intermittent sulphadoxine-pyrimethamine to prevent severe anaemia secondary to malaria in pregnancy: a randomised placebo-controlled trial C E Shulman, E K Dorman, F Cutts, K Kawuondo, J N Bulmer, N Peshu, K Marsh
Summary Background In areas of endemic transmission, malaria in pregnancy is associated with severe maternal anaemia and low-birthweight babies. We studied the efficacy of intermittent treatment doses of sulphadoxine-pyrimethamine in preventing malaria and severe anaemia in pregnancy in a double-blind placebo-controlled trial among primigravid women living in Kilifi District, Kenya. Methods Between January, 1996, and April, 1997, 1264 primigravid women were recruited when they attended for antenatal care, and randomly assigned sulphadoxinepyrimethamine (640) or placebo (624). Women received one, two, or three doses of study medication depending on the duration of gestation at enrolment. Primary outcome measures were severe anaemia (haemoglobin <8 g/dL) and malaria parasitaemia, assessed at 34 weeks of gestation. Analyses were based on intention to treat among women who had study blood tests at 34 weeks. Findings 30 (5·3%) of 567 women in the sulphadoxinepyrimethamine group and 199 (35·3%) of 564 in the placebo group had peripheral parasitaemia (protective efficacy 85% [95% CI 78–90], p<0·0001). 82 (14·5%) and 134 (23·7%) had severe anaemia (protective efficacy 39% [22–52], p<0·0001). Even women who booked late and received only one dose of sulphadoxine-pyrimethamine benefited significantly from the intervention. The effects were seen both in women who owned insecticide-treated bednets and in women who did not. Interpretation Intermittent presumptive treatment with sulphadoxine-pyrimethamine is an effective, practicable strategy to decrease the risk of severe anaemia in primigravidae living in malarious areas.
Lancet 1999; 353: 632–36
Introduction In areas where infection with Plasmodium falciparum is endemic, immunity to malaria develops over the first few
years of life, and older children and adults rarely suffer severe complications of infection. An exception is during pregnancy, when risks of maternal anaemia1 and low birthweight2 are increased by malaria infection.3 The risks are greatest in the first pregnancy and decrease with increasing gravidity.3,4 Because infection is symptomless in most cases,5 treatment of symptomatic episodes only will miss the majority of infections. Many peripheralblood slides are negative, despite the presence of sequestered parasites in the placenta.6,7 Previous trials have shown that malaria chemoprophylaxis given to primigravidae is associated with less placental parasitaemia, a lower proportion of low-birthweight babies,2,8 and higher mean maternal haemoglobin;9 such prophylaxis may prevent severe haemolytic anaemia.10,11 Chloroquine has been the main drug used in sub-Saharan Africa. However, there are now very high rates of drug resistance,2 and poor compliance is common.12 Many countries no longer recommend chloroquine prophylaxis during pregnancy. Consequently, new preventive strategies have been sought.13 Possible strategies include other forms of malaria control such as insecticide-treated bednets or a different drug regimen. Two studies of insecticide-treated bednets carried out in areas of low malaria transmission14,15 showed some reduction in anaemia during pregnancy when nets were used. In areas of higher transmission, however, there was no benefit on maternal health,7 despite a significant decrease in childhood mortality and morbidity.16,17 We studied whether, in primigravidae, a regimen of between one and three presumptive treatment doses of sulphadoxine-pyrimethamine in the second trimester and early third trimester of pregnancy would prevent severe maternal anaemia in late pregnancy. Prevention of this disorder would be likely to have significant benefit for the mother and child.3 Previous work in Malawi showed that such a regimen had a significant effect on placental parasitaemia.18 We also compared the effects of this intervention in a population that had previously received insecticide-treated bednets and a population that had not.
Participants and methods London School of Hygiene and Tropical Medicine, London, UK (C E Shulman MRCGP, F Cutts MD); Kenya Medical Research Institute (KEMRI), Centre for Geographical Medicine Research Coast, Kilifi, Kenya (C E Shulman, E K Dorman MRCOG, K Kawuondo HND, N Peshu MB, Prof K Marsh FRCP); University of Oxford, Institute of Molecular Medicine, John Radcliffe Hospital, Oxford (E K Dorman, K Marsh); and Department of Pathology, University of Newcastle, Royal Victoria Infirmary, Newcastle upon Tyne, UK (J N Bulmer MRCPath) Correspondence to: Dr C E Shulman, Maternal and Child Epidemiology Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK (e-mail: [email protected])
632
Study population and area The study took place in 1996–97 in Kilifi, a mainly rural district on the coast of Kenya. Transmission of P falciparum is perennial with two seasonal peaks (June–August and November–December).19,20 The transmission intensity is higher in the south of the district than in the north. The point prevalence of infection in 1–9-year-old children was 74% in Kilifi south (hyperholoendemic) and 49% in Kilifi north (mesohyperendemic).21 A previous study in this area1 showed that pregnant women had a high prevalence of severe anaemia; the prevalence was twice as high in primigravid as in multigravid women. Iron deficiency was very common in this population (75% of women had serum ferritin <12 μg/L), as was hookworm
THE LANCET • Vol 353 • February 20, 1999
ARTICLES
infestation (75% of women had hookworm eggs in their faeces). In primigravidae, the presence of peripheral parasitaemia was the main determinant of severe anaemia. Of the study population, 80% lived in rural areas—roughly half of these were from Kilifi south and half from Kilifi north. Most of the rural population from Kilifi north had taken part in a previous community randomised study of insecticide-treated bednets.17 Repeat dipping of bednets in insecticide continued every 6 months throughout the period of our study at several sites within the original bednet study area.
Design To be eligible, a woman had to be attending the antenatal clinic at Vipingo Health Centre (Kilifi south) or Kilifi District Hospital (Kilifi north) and primigravid with a singleton pregnancy of between 16 and 30 weeks’ gestation. Exclusion criteria were multiple pregnancy, haemoglobin below 6 g/dL, severe preeclampsia, a clinical diagnosis of malaria, an intrauterine fetal death, or a previous suspected reaction to a sulpha drug. Written informed consent was obtained from all participants. Ethical clearance was granted by the Kenyan Medical Research Institute/National ethical review committee, and the London School of Hygiene and Tropical Medicine. A safety monitoring committee was established. Women were individually randomised to sulphadoxinepyrimethamine or placebo, identical in appearance and taste. Participants were assigned unique identification numbers sequentially. All identification numbers had been randomly allocated to a number between zero and nine, in blocks of ten. These numbers corresponded to ten bottles, five of which were randomly allocated to contain sulphadoxine-pyrimethamine and five placebo. Questionnaires were premarked with this unique identification number and the bottle number. The code relating bottle numbers to their contents was retained by a statistician and clinician at the research unit, who were not involved in the study. The investigators remained unaware of the assigned group until after data collection and editing were complete. The target number of doses of sulphadoxine-pyrimethamine or placebo depended on gestation at recruitment—three doses for women recruited at 16–19 weeks of gestation; two for those recruited at 20–26 weeks; and one for those recruited at 27–30 weeks. All doses were given under observation at clinic visits. Women received 30 days’ supply of ferrous sulphate at each visit and continued to receive routine antenatal care from the hospital midwives. At recruitment a standard questionnaire was administered on pregnancy history, history of illnesses, drug use, and educational attainment of the woman and her partner. Duration of gestation was assessed by history of last menstrual period and ultrasonography. At each subsequent visit, women were asked about any possible drug reactions, and further doses were withheld if they reported skin or systemic reactions thought to be related to the study drugs. Information was recorded on antimalarial medication from public, private, or traditional sectors and on whether the woman’s family owned bednets. At about 34 weeks of gestation, women completed the antenatal component of the study with a blood test for haemoglobin concentration, malaria slide, and haemoglobin electrophoresis. HIV testing was done anonymously with linkage after deletion of personal identifiers. At the time of this blood test, all women who were anaemic were treated and followed up as appropriate, to reduce the likelihood of severe anaemia at the time of delivery. Women with haemoglobin below 10 g/dL were given folic acid and ferrous sulphate. Women with haemoglobin below 7 g/dL or with parasitaemia were given sulphadoxinepyrimethamine and those with haemoglobin below 6 g/dL were admitted to hospital. If study nurses suspected severe anaemia before 34 weeks’ gestation, haemoglobin was measured immediately. If the concentration was less than 7 g/dL, other investigations were done and treatment given, as described for the 34-week visit.
THE LANCET • Vol 353 • February 20, 1999
Full blood counts were done on a model M530 Coulter Counter. Haemoglobinopathies were identified by haemoglobin electrophoresis. Thick and thin blood films were prepared by standard giemsa staining, with counting of parasites per 200 white cells and calculation of counts per μL from the white-cell count. 100 high-power fields were read to confirm a negative film. Evidence of HIV infection was sought by GACPAT (Public Health Laboratory Service, London, UK).22 For women who gave birth in hospital, thick blood smears were taken at the time of delivery from the maternal surface of the placenta and examined for parasites and pigment by standard giemsa staining. Placental biopsy samples were prepared and examined by two people as described previously.23,24 Women were seen at least 4 weeks after delivery, so that stillbirths, neonatal deaths, maternal deaths, and maternal and neonatal morbidity could be ascertained.
Statistics Data entry, validation, and cleaning were done with DBase IV. Statistical analysis was carried out with STATA (release 4·0) by intention to treat for women who had a blood test at 34 weeks. The primary outcome was haemoglobin concentration below 8 g/dL at 34 weeks of gestation. Protective efficacy of sulphadoxine-pyrimethamine compared with placebo for specific endpoints was calculated from the risk ratios for those endpoints: protective efficacy=1003(1–risk ratio). Subgroup analyses were done on number of doses allocated and on areas of residence. For each potential confounder (educational status of woman and her partner, ownership of radio, type of house, marital status, number of other wives, sickle-cell trait, HIV infection, antimalarials received outside of the study protocol, area of residence, and bednet ownership), an adjusted risk ratio estimated by the Mantel-Haenszel technique was compared with the unadjusted risk ratio. For all such analyses, crude and adjusted risk ratios showed no evidence of confounding, hence the risk ratios presented are unadjusted. The sample size was based on that required to measure a 50% reduction in the proportion of women with haemoglobin below 8 g/dL with 80% power at 5% significance, separately for the population in the bednet-study area and the population not in that study area. We aimed to recruit 500 women from each area. However, a higher proportion of women than expected were not eligible because they were more than 30 weeks pregnant at first attendance, so we recruited only 429 women from the population from the bednet-study area, of whom 70% still used an insecticide-treated bednet. 1601 women screened 337 not eligible 268 ⬎30 weeks' gestation 46 severe anaemia or clinical malaria 11 severe pre-eclampsia or fetal death 10 refused 2 possible reaction to sulpha drugs 1264 women randomised
624 assigned placebo
59 had no blood test during third trimester 565 had blood test during third trimester
640 assigned sulphadoxine-pyrimethamine 73 had no blood test during third trimester 567 had blood test during third trimester
Trial profile
633
ARTICLES
Sulphadoxinepyrimethamine (n=640)
Placebo (n=623)*
Demography Ethnic group–Mijikenda Married
565 (88·3%) 538 (84·1%)
560 (89·9%) 506 (81·2%)
Educational status Able to read a letter No education Incomplete primary Primary complete Secondary complete
383 (59·8%) 206 (32·2%) 248 (38·8%) 111 (17·3%) 75 (11·7%)
370 (59·4%) 211 (33·9%) 243 (39·0%) 109 (17·5%) 60 (9·6%)
Educational status of husband/partner† Able to read Primary complete Secondary complete
557 (93·5%) 422 (70·8%) 161 (27·0%)
541 (95·3%) 417 (73·4%) 158 (27·8%)
358 (55·9%) 341 (53·3%) 144 (22·5%)
336 (53·9%) 334 (53·6%) 128 (20·6%)
81 (14·5%) 29 (5·1%) 23·0
80 (14·3%) 33 (5·8%) 23·1
Socioeconomic status Owns radio Latrine in compound House walls made of coral/stone (ie, not mud) Medical Sickle-cell trait‡ HIV positive§ Mean duration of gestation at recruitment (weeks)
Medication taken during pregnancy, before enrolment Sulphadoxine-pyrimethamine 18 (2·8%) Chloroquine treatment 36 (5·6%) Chloroquine prophylaxis 27 (4·2%)
12 (1·9%) 38 (6·1%) 25 (4·0%)
*Information available for only 623 of 624 women recruited. †n=596 for sulphadoxine-pyrimethamine group; n=568 for placebo group. ‡n=557 for sulphadoxine-pyrimethamine group; n=561 for placebo group. §n=571 for sulphadoxine-pyrimethamine group; n=565 for placebo group.
Table 1: Characteristics of intervention and control groups
Results Study population 1601 primigravid women attended clinic for the first time during the recruitment period (figure). 337 were not eligible. Of the 1264 women assigned sulphadoxinepyrimethamine or placebo, 233 (18%) were allocated three doses, 811 (64%) two doses, and 220 (17%) one dose. The study groups were similar in terms of ethnic group, education, socioeconomic status, gestation at recruitment, HIV status, frequency of sickle-cell trait, and drug history during the pregnancy (table 1). Efficacy Of 1264 women with a singleton pregnancy enrolled within the trial, 1132 (90%) had a study blood test: 1054 after 32 weeks’ gestation, 66 at 28–32 weeks’ gestation, and 12 at less than 28 weeks’ gestation. There was no
All participants
Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
82/567 (14·5%)
134/565 (23·7%)
39 (22 to 52)
<0·0001
By doses allocated One 12/99 (12%) Two 50/350 (14%) Three 20/118 (17%) Residence Rural south Rural north Total With ITBN No ITBN Urban
24/99 (24%) 88/378 (23%) 22/88 (25%)
50 (6 to 73) 39 (16 to 55) 32 (216 to 60)
0·027 0·002 0·16
36/198 (18%)
61/211 (29%)
37 (10 to 56)
0·011
26/243 (11%) 12/146 (8%) 14/97 (14%) 20/126 (16%)
48/236 (20%) 27/148 (18%) 21/88 (24%) 25/117 (21%)
47 (18 to 66) 55 (15 to 76) 40 (211 to 67) 26 (226 to 56)
0·004 0·011 0·102 0·271
difference in duration of gestation at blood test between the groups (p=0·44), and the reasons for early blood testing were similar in both groups. 11 women from the sulphadoxine-pyrimethamine group and 17 from the placebo group had the study blood test early owing to severe anaemia or clinical malaria. The most common reason for not having a study blood test was delivery before 35 weeks’ gestation (86 [65%] of 132). 118 (9·3%) women (59 from each group) did not take the number of doses that they had been allocated. There were no substantial differences between the groups in the reasons for non-completion of the assigned course, and no serious adverse reactions to the medication. Four women in the sulphadoxine-pyrimethamine group and three in the placebo group had study medication suspended because of suspected minor drug reaction. Among the 1132 women who had the third-trimester study blood test, a significantly higher proportion of the placebo group than of the sulphadoxine-pyrimethamine group were anaemic (haemoglobin ∂11 g/dL: 460 [81%] vs 431 [76%], p=0·026). Presumptive treatment with sulphadoxine-pyrimethamine also led to a lower rate of severe anaemia (haemoglobin <8 g/dL) with an overall protective efficacy of 39% (95% CI 22–52) and an absolute risk difference of 9·2% (p<0·0001, table 2). Efficacy was high and was significant even if only one dose of medication was given (p=0·027). The overall effect was similar in both rural populations, including the women who still had insecticide-treated bednets, though urban women did not seem to derive as much benefit from the intervention as rural women did (table 2). In some of the subsets, however, the sample sizes were small, so the statistical power was low. We also examined the effect of the intervention by place of recruitment (Vipingo or Kilifi District Hospital). There was less anaemia with sulphadoxinepyrimethamine than with placebo in all subgroups except women who gave an address from Kilifi south, but who attended the clinic at Kilifi District Hospital. In that subgroup, 16 (22%) of 73 women in the sulphadoxinepyrimethamine group and eight (11%) of 71 in the placebo group had severe anaemia (p=0·07). By contrast, the effect of the intervention in the rural population of Kilifi south who went to Vipingo clinic was large: 20 (16%) of 125 in the sulphadoxine-pyrimethamine group versus 53 (39%) of 137 in the placebo group had severe anaemia, giving a protective efficacy of 59% (p<0·0001). Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
All participants
30/567 (5·3%)
199/564 (35·3%)
85 (78–90)
<0·0001
By doses allocated One Two Three
4/99 (4%) 19/350 (5%) 7/118 (6%)
27/99 (27%) 134/377 (35%) 38/88 (43%)
85 (59–95) 85 (76–90) 86 (71–94)
<0·0001 <0·0001 <0·0001
21/198 (11%)
100/210 (48%)
78 (66–85)
<0·0001
4/243 (2%) 3/146 (2%) 1/97 (1%) 5/126 (4%)
66/236 (28%) 36/148 (24%) 30/88 (34%) 33/117 (28%)
94 (84–98) 92 (73–97) 97 (78–99) 86 (65–94)
<0·0001 <0·0001 <0·0001 <0·0001
Residence Rural south Rural north Total With ITBN No ITBN Urban
For placebo group; information on residence available for only 564 of 565 with antenatal haemoglobin measurements. ITBN=insecticide-treated bednet.
For placebo group; information on residence available for only 563 of 564 with antenatal blood film. ITBN=insecticide-treated bednet.
Table 2: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on proportion of women with severe anaemia
Table 3: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on proportion of women with peripheral parasitaemia
634
THE LANCET • Vol 353 • February 20, 1999
ARTICLES
Miscarriages* Stillbirths Perinatal deaths Neonatal deaths
Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
4/630 (0·6%) 24/626 (3·8%) 39/626 (6·2%) 19/602 (3·2%)
4/615 (0·7%) 26/611 (4·3%) 49/611 (8·0%) 30/585 (5·1%)
2·7 (2287 to 76) 0·4 (22 to 3) 22 (217 to 48) 38 (28 to 65)
0·97 0·71 0·22 0·09
*Delivery at less than 24 weeks’ gestation.
Table 4: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on postnatal outcome
Mean haemoglobin concentration was significantly higher in the sulphadoxine-pyrimethamine group than in the placebo group (mean 9·7 g/dL [SD 1·76; range 2·7–14·6] vs 9·3 g/dL [1·85; 3·4–14·5], p<0·0001). 1131 women had peripheral blood film results. Among these women, malaria parasitaemia was less common in the sulphadoxine-pyrimethamine group than in the placebo group, with a protective efficacy of 85% (95% CI 78–90, p<0·0001); table 3). The effect was similar in all subgroups. There was little difference between the groups in reported treatment-seeking behaviour or medication during the study. Six women from the sulphadoxinepyrimethamine group and eight from the placebo group reported taking sulphadoxine-pyrimethamine, outside the study protocol, and 69 and 61 women, respectively, reported taking chloroquine. Five of 29 HIV-infected women in the sulphadoxinepyrimethamine group compared with nine of 33 from the placebo group had haemoglobin less than 8 g/dL, and three and ten, respectively, had peripheral parasitaemia. However, the differences were not significant. Among the women who gave birth in hospital, 16 (7·8%) of 205 in the sulphadoxine-pyrimethamine group and 29 (14·8%) of 196 in the placebo group had a placental smear positive for malaria (protective efficacy 47% [6–70], p=0·27); in 93 (45·4%) of 205 and 69 (34·2%) of 202, respectively, the placenta was negative for malaria on histology (protective efficacy 17% [3–29], p=0·021).
Postnatal follow-up Information on birth outcome was obtained from 1245 (98·5%) women. There were fewer stillbirths and neonatal deaths in the sulphadoxine-pyrimethamine group than in the placebo group, consistent with a protective efficacy of 22% for perinatal deaths and 38% for neonatal mortality, though these differences were not significant (table 4). There was no difference in the proportion of women delivering prematurely or in the mean duration of gestation at delivery between the groups. Of five babies who died and were reported to have been jaundiced, three were born to mothers who had not received sulphadoxine-pyrimethamine during pregnancy and two were born to mothers who had received the drugs more than 6 weeks before delivery. Five women died, four in the placebo group (one from chorioamnionitis at 30 weeks’ gestation, one with a history of acute abdominal pain antenatally, and two as a result of obstructed labour) and one in the sulphadoxinepyrimethamine group (with a history of severe pallor and acute diarrhoea at 30 weeks’ gestation). Of these women, the four who were tested for HIV were negative.
Discussion In this double-blind randomised controlled trial, presumptive treatment with sulphadoxine-pyrimethamine
THE LANCET • Vol 353 • February 20, 1999
greatly reduced the frequency of severe anaemia and was acceptable to women. Since the majority of primigravid women in Kenya attend for antenatal care (estimates are 95% for women attending antenatal clinics at least once in Kilifi District)25 and sulphadoxine-pyrimethamine is cheap (each dose costs US$0·17), this regimen should be practicable. The finding of a significant benefit from the intervention in women who attended late in pregnancy and so were assigned only one dose (table 2), suggests that even these women can benefit from the intervention and that health-workers should try as hard to give late attenders a single dose as to provide early attenders with two or three doses. However, the findings cannot be interpreted as meaning that one dose is sufficient treatment for all women. Women may develop severe lifethreatening anaemia secondary to malaria early in the second trimester, so presumptive treatment needs to be started as early as possible after 16 weeks of gestation. Furthermore, there may be differences between women who book for antenatal clinic late and women coming earlier in pregnancy, for example, in the proportion with illness. This regimen has the advantage over chloroquine that it is given to women when they attend antenatal clinics, thus helping to avoid problems of compliance. We did not detect any serious side-effects in mothers or infants, and there was no difference in minor adverse events between the groups. There is a small risk of severe adverse reactions to sulpha drugs,26 and a theoretical risk of kernicterus if these drugs are used within a short period before delivery.27 However, in an area with endemic malarial transmission, and much severe anaemia attributable to malaria in pregnancy, the benefits of this regimen outweigh these potential risks. No increased risk of kernicterus has been seen in this or other studies of sulphadoxine-pyrimethamine use in pregnancy.18,28 In our study, however, only 42 women received the drugs less than 14 days before delivery; the mean time between the last dose and delivery was 56 days. Although these results are reassuring, surveillance systems remain important for safety monitoring, if this regimen becomes routine. The rural population from the north had less anaemia than the southern population (table 2). These are geographically distinct populations, so there could be several reasons for this difference, including intensity of malaria transmission,21 variations in nutritional status, and variations in hookworm prevalence. The absolute difference in rates of anaemia between intervention and placebo groups was similar in all the rural subpopulations, including the population with insecticidetreated bednets. This finding confirms previous reports that bednets alone do not prevent the adverse outcomes of malaria in pregnancy.7 It also suggests that treatment with sulphadoxine-pyrimethamine is effective in a range of transmission settings, though was not as effective in the urban populations as in the rural populations. The only group of women who appeared to derive no benefit from the intervention were women recruited at Kilifi District Hospital who gave an address in Kilifi south. The proportion with anaemia among these women was similar to that in the urban population; these women may have been residing temporarily in the urban areas of Kilifi north. Severe anaemia in pregnancy has been reported as the main cause of up to 20% of maternal deaths in some hospital series in sub-Saharan Africa29,30 and 11–13% in community-based studies.31,32 In Kilifi District Hospital, 635
ARTICLES
severe anaemia was judged to be the primary factor in ten of 43 deaths over a 3-year period (EKD, CES; unpublished). Severe anaemia is also associated with substantial maternal morbidity. Malaria infection and severe maternal anaemia are also risk factors for low birthweight,3 one of the main risk factors for infant mortality.33 On the assumption that without treatment, anaemia at 34 weeks’ gestation will not resolve spontaneously, a decrease in the rate of severe anaemia in the third trimester is likely to have a great public-health impact. Malaria is an under-recognised cause of severe anaemia in areas of endemic transmission, since it is generally asymptomatic and not associated with peripheral parasitaemia. The delay in the development of an effective regimen to prevent malaria in pregnant women since the emergence of widespread chloroquine resistance has resulted in much morbidity and mortality. Unlike previous policies for pregnant women, this regimen does not involve continuous prophylaxis and is therefore unlikely to contribute significantly to the emergence of drug resistance. Use in pregnant women will form only a small part of total use. Nonetheless, the likely emergence of resistance to sulphadoxinepyrimethamine is a concern, and research to identify ways of preventing such resistance and of finding alternative treatments should continue in parallel with implementation of policy. Contributors Caroline Shulman designed the study, coordinated and supervised data collection, did the analysis, and wrote the paper. Ed Dorman carried out all ultrasound scans, supervised data collection, and was involved in the clinical management. Felicity Cutts contributed to the study and advised on the analysis and technical and practical issues. Ken Kawuondo undertook and supervised day-to-day laboratory work. Judith Bulmer undertook the histological examinations of the placenta. Norbert Peshu contributed to the study design and advised on practical issues. Kevin Marsh contributed to the study design and advised on the analysis and technical and practical issues. All investigators contributed to the writing of the paper.
Acknowledgments This paper is published with the permission of the Director of KEMRI. The work was supported by the UK Department for International Development and KEMRI. We thank Peter Smith, Brian Greenwood, Bernard Brabin, and Cathy Waruiru, who were members of the safety monitoring committee; Bob Snow for his advice, all at the Kilifi research unit, particularly Brett Lowe who was responsible for managing the laboratory work; midwives Jane Mwendwa, Judith Peshu, and Ann Muhoro; the study fieldworkers; Oona Campbell and Tom Marshall for their advice and valuable comments on the paper; and John Parry (Hepatitis and Retrovirus Laboratory, Public Health Laboratory Service, London) who assisted in the supply of kits and advice on HIV testing. KM is a Wellcome Trust senior research fellow (631342).
8
9
10
11 12
13
14
15
16
17
18
19
20
21
22
23
24
25
References 1
2
3 4
5
6
7
Shulman C, Graham W, Jilo H, et al. Malaria as an important cause of anaemia in primigravidae: evidence from a district hospital in coastal Kenya. Trans R Soc Trop Med Hyg 1996; 90: 535–39. Steketee RW, Wirima JJ, Hightower AW, et al. The effect of malaria and malaria prevention in pregnancy on offspring birthweight, prematurity, and intrauterine growth retardation in rural Malawi. Am J Trop Med Hyg 1996; 55 (suppl 1): 33–44. Brabin B. The risks and severity of malaria in pregnant women. Appl Field Research in Malaria reports WHO 1991; no1. Steketee RW, Breman JG. Malaria infection in pregnant women in Zaire: the effects and potential for intervention. Ann Trop Med Parasitol 1988; 82: 113–20. McDermott JM, Heymann DL, Wirima JJ, et al. Efficacy of chemoprophylaxis in preventing Plasmodium falciparum parasitaemia and placental infection in pregnant women in Malawi. Trans R Soc Trop Med Hyg 1988; 82: 520–23. Descowitz RS, Alpers MP. Placental Plasmodium falciparum parasitaemia in East Sepik women of different parity. Ann Trop Med Parasitol 1992; 86: 95–102. Shulman C, Dorman E, Talisuna A, et al. A community randomised controlled trial of insecticide treated bednets for the prevention of
636
26
27
28 29 30 31
32
33
malaria and anaemia among primigravid women on the Kenyan Coast. Trop Med Int Health 1998; 3: 197–204. Mutabingwa TK, Malle LN, de Geus A, Oosting J. Malaria chemosuppression in pregnancy II: its effect on maternal haemoblobin levels, placental malaria and birth weight. Trop Geogr Med 1993; 45: 49–55. Greenwood BM, Greenwood AM, Snow RW, Byass P, Bennett S, Hatib-N’Jie AB. The effects of malaria chemoprophylaxis given by traditional birth attendants on the course and outcome of pregnancy. Trans R Soc Trop Med Hyg 1989; 83: 589–94. Fleming AF, Ghatoura GBS. The prevention of anaemia in pregnancy in primigravidae in the guinea savanna of Nigeria. Ann Trop Med Parasitol 1986; 80: 211–33. Gilles HM, Lawson JB, Sibelas M, Voller A, Allan N. Malaria, anaemia, and pregnancy. Ann Trop Med Parasitol 1969; 63: 245–63. Heyman DL, Steketee RW, Wirima JJ, McFarland DA, Khoromana CO, Campbell CC. Antenatal chloroquine prophylaxis in Malawi: chloroquine resistance, compliance, protective efficacy and cost. Trans R Soc Trop Med Hyg 1990; 84: 496–98. Gülmezoglu AM, Garner P. Malaria in pregnancy in endemic areas. (Cochrane Review) In: The Cochrane Library : Issue 4, 1998. Oxford: Update Software. Dolan G, ter-Kuile FO, Jacoutot V, et al. Bed nets for the prevention of malaria and anaemia in pregnancy. Trans R Soc Trop Med Hyg 1993; 87: 620–26. D’Alessandro U, Langerock P, Bennett S, Francis N, Cham K, Greenwood BM. The impact of a national impregnated bed net programme on the outcome of pregnancy in primigravidae in The Gambia. Trans R Soc Trop Med Hyg 1996; 90: 487–92. Alonso PL, Lindsay SW, Armstrong JRM, et al. The effect of insecticide-treated bed nets on the mortality of Gambian children. Lancet 1991; 337: 1499–502. Nevill CG, Some ES, Mung’ala VO, et al. Insecticide treated bednets reduce mortality and severe morbidity from malaria among children on the Kenyan coast. Trop Med Int Health 1996; 1: 139–46. Schultz LJ, Steketee RW, Macheso A, Kazembe P, Chitsulo L, Wirima JJ. The efficacy of antimalarial regimens containing sulfadoxine-pyrimethamine and/or chloroquine in preventing peripheral and placental plasmodium falciparum infection among pregnant women in Malawi. Am J Trop Med Hyg 1994; 51: 515–22. Mbogo CNM, Snow RW, Khamala CPM, et al. Relationship between Plasmodium falciparum transmission by vector populations and the incidence of severe disease at nine sites on the Kenyan Coast. Am J Trop Med Hyg 1995; 52: 201–06. Snow RW, Schellenberg JR, Peshu N, et al. Periodicity and space-time clustering of severe childhood malaria on the coast of Kenya. Trans R Soc Trop Med Hyg 1993; 87: 386–90. Snow RW, Omumbo JA, Lowe B, et al. Relation between severe malaria morbidity in children and level of Plasmodium falciparum transmission in Africa. Lancet 1997; 349: 1650–54. Parry JV, Mortimer PP. An immunoglobulin G antibody capture particle-adherence test (GACPAT) for antibody to HIV-1 and HTLV1 that allows economical large-scale screening. AIDS 1989; 3: 173–76. Bulmer JN, Rasheed FN, Francis N, Morrison L, Greenwood BM. Placental malaria I: pathological classification. Histopathology 1993; 22: 211–18. Bulmer JN, Rasheed FN, Morrison L, Francis N, Greenwood BM. Placental malaria II: a semi-quantitative investigation of the pathological features. Histopathology 1993; 22: 219–25. Kirumbi LW, Wassuna A, Nabwani G, Mpanju-Shumbusho W. Prevalence of and factors associated with perinatal morbidity and mortality in Kirinyaga and Kilifi Districts in Kenya. Project report, 1996. Bjorkman A, Phillips-Howard PA. Adverse reactions to sulpha drugs: implications for malaria chemotherapy. Bull World Health Organ 1991; 69: 297–304. Walker PC. Neonatal bilirubin toxicity: a review of kernicterus and the implications of drug-induced bilirubin displacement. Clin Pharmacokinet 1987; 13: 26–50. Phillips-Howard PA, Wood D. The safety of antimalarial drugs in pregnancy. Drug Saf 1996; 14: 131–45. Armon PJ. Maternal deaths in Tanzania. Trans R Soc Trop Med Hyg 1979; 73: 284–88. Mtimavalye LAR, Lisasi D, Ntuyabaliwe WK. Maternal mortality in Dar es Salaam, Tanzania. East Afr Med J 1980; 57: 111–18. Boerma JT, Mati JKG. Identifying maternal mortality through networking: results from coastal Kenya. Stud Fam Plann 1989; 20: 245–53. MacLeod J, Rhode R. Retrospective follow-up of maternal deaths and their associated risk factors in a rural district of Tanzania. Trop Med Int Health 1998; 3: 130–37. McCormick MC. The contribution of low birth weight to infant mortality and childhood mortality. N Engl J Med 1985; 31: 82–90.
THE LANCET • Vol 353 • February 20, 1999
Intermittent sulphadoxine-pyrimethamine to prevent severe anaemia secondary to malaria in pregnancy: a randomised placebo-controlled trial C E Shulman, E K Dorman, F Cutts, K Kawuondo, J N Bulmer, N Peshu, K Marsh
Summary Background In areas of endemic transmission, malaria in pregnancy is associated with severe maternal anaemia and low-birthweight babies. We studied the efficacy of intermittent treatment doses of sulphadoxine-pyrimethamine in preventing malaria and severe anaemia in pregnancy in a double-blind placebo-controlled trial among primigravid women living in Kilifi District, Kenya. Methods Between January, 1996, and April, 1997, 1264 primigravid women were recruited when they attended for antenatal care, and randomly assigned sulphadoxinepyrimethamine (640) or placebo (624). Women received one, two, or three doses of study medication depending on the duration of gestation at enrolment. Primary outcome measures were severe anaemia (haemoglobin <8 g/dL) and malaria parasitaemia, assessed at 34 weeks of gestation. Analyses were based on intention to treat among women who had study blood tests at 34 weeks. Findings 30 (5·3%) of 567 women in the sulphadoxinepyrimethamine group and 199 (35·3%) of 564 in the placebo group had peripheral parasitaemia (protective efficacy 85% [95% CI 78–90], p<0·0001). 82 (14·5%) and 134 (23·7%) had severe anaemia (protective efficacy 39% [22–52], p<0·0001). Even women who booked late and received only one dose of sulphadoxine-pyrimethamine benefited significantly from the intervention. The effects were seen both in women who owned insecticide-treated bednets and in women who did not. Interpretation Intermittent presumptive treatment with sulphadoxine-pyrimethamine is an effective, practicable strategy to decrease the risk of severe anaemia in primigravidae living in malarious areas.
Lancet 1999; 353: 632–36
Introduction In areas where infection with Plasmodium falciparum is endemic, immunity to malaria develops over the first few
years of life, and older children and adults rarely suffer severe complications of infection. An exception is during pregnancy, when risks of maternal anaemia1 and low birthweight2 are increased by malaria infection.3 The risks are greatest in the first pregnancy and decrease with increasing gravidity.3,4 Because infection is symptomless in most cases,5 treatment of symptomatic episodes only will miss the majority of infections. Many peripheralblood slides are negative, despite the presence of sequestered parasites in the placenta.6,7 Previous trials have shown that malaria chemoprophylaxis given to primigravidae is associated with less placental parasitaemia, a lower proportion of low-birthweight babies,2,8 and higher mean maternal haemoglobin;9 such prophylaxis may prevent severe haemolytic anaemia.10,11 Chloroquine has been the main drug used in sub-Saharan Africa. However, there are now very high rates of drug resistance,2 and poor compliance is common.12 Many countries no longer recommend chloroquine prophylaxis during pregnancy. Consequently, new preventive strategies have been sought.13 Possible strategies include other forms of malaria control such as insecticide-treated bednets or a different drug regimen. Two studies of insecticide-treated bednets carried out in areas of low malaria transmission14,15 showed some reduction in anaemia during pregnancy when nets were used. In areas of higher transmission, however, there was no benefit on maternal health,7 despite a significant decrease in childhood mortality and morbidity.16,17 We studied whether, in primigravidae, a regimen of between one and three presumptive treatment doses of sulphadoxine-pyrimethamine in the second trimester and early third trimester of pregnancy would prevent severe maternal anaemia in late pregnancy. Prevention of this disorder would be likely to have significant benefit for the mother and child.3 Previous work in Malawi showed that such a regimen had a significant effect on placental parasitaemia.18 We also compared the effects of this intervention in a population that had previously received insecticide-treated bednets and a population that had not.
Participants and methods London School of Hygiene and Tropical Medicine, London, UK (C E Shulman MRCGP, F Cutts MD); Kenya Medical Research Institute (KEMRI), Centre for Geographical Medicine Research Coast, Kilifi, Kenya (C E Shulman, E K Dorman MRCOG, K Kawuondo HND, N Peshu MB, Prof K Marsh FRCP); University of Oxford, Institute of Molecular Medicine, John Radcliffe Hospital, Oxford (E K Dorman, K Marsh); and Department of Pathology, University of Newcastle, Royal Victoria Infirmary, Newcastle upon Tyne, UK (J N Bulmer MRCPath) Correspondence to: Dr C E Shulman, Maternal and Child Epidemiology Unit, Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London WC1E 7HT, UK (e-mail: [email protected])
632
Study population and area The study took place in 1996–97 in Kilifi, a mainly rural district on the coast of Kenya. Transmission of P falciparum is perennial with two seasonal peaks (June–August and November–December).19,20 The transmission intensity is higher in the south of the district than in the north. The point prevalence of infection in 1–9-year-old children was 74% in Kilifi south (hyperholoendemic) and 49% in Kilifi north (mesohyperendemic).21 A previous study in this area1 showed that pregnant women had a high prevalence of severe anaemia; the prevalence was twice as high in primigravid as in multigravid women. Iron deficiency was very common in this population (75% of women had serum ferritin <12 μg/L), as was hookworm
THE LANCET • Vol 353 • February 20, 1999
ARTICLES
infestation (75% of women had hookworm eggs in their faeces). In primigravidae, the presence of peripheral parasitaemia was the main determinant of severe anaemia. Of the study population, 80% lived in rural areas—roughly half of these were from Kilifi south and half from Kilifi north. Most of the rural population from Kilifi north had taken part in a previous community randomised study of insecticide-treated bednets.17 Repeat dipping of bednets in insecticide continued every 6 months throughout the period of our study at several sites within the original bednet study area.
Design To be eligible, a woman had to be attending the antenatal clinic at Vipingo Health Centre (Kilifi south) or Kilifi District Hospital (Kilifi north) and primigravid with a singleton pregnancy of between 16 and 30 weeks’ gestation. Exclusion criteria were multiple pregnancy, haemoglobin below 6 g/dL, severe preeclampsia, a clinical diagnosis of malaria, an intrauterine fetal death, or a previous suspected reaction to a sulpha drug. Written informed consent was obtained from all participants. Ethical clearance was granted by the Kenyan Medical Research Institute/National ethical review committee, and the London School of Hygiene and Tropical Medicine. A safety monitoring committee was established. Women were individually randomised to sulphadoxinepyrimethamine or placebo, identical in appearance and taste. Participants were assigned unique identification numbers sequentially. All identification numbers had been randomly allocated to a number between zero and nine, in blocks of ten. These numbers corresponded to ten bottles, five of which were randomly allocated to contain sulphadoxine-pyrimethamine and five placebo. Questionnaires were premarked with this unique identification number and the bottle number. The code relating bottle numbers to their contents was retained by a statistician and clinician at the research unit, who were not involved in the study. The investigators remained unaware of the assigned group until after data collection and editing were complete. The target number of doses of sulphadoxine-pyrimethamine or placebo depended on gestation at recruitment—three doses for women recruited at 16–19 weeks of gestation; two for those recruited at 20–26 weeks; and one for those recruited at 27–30 weeks. All doses were given under observation at clinic visits. Women received 30 days’ supply of ferrous sulphate at each visit and continued to receive routine antenatal care from the hospital midwives. At recruitment a standard questionnaire was administered on pregnancy history, history of illnesses, drug use, and educational attainment of the woman and her partner. Duration of gestation was assessed by history of last menstrual period and ultrasonography. At each subsequent visit, women were asked about any possible drug reactions, and further doses were withheld if they reported skin or systemic reactions thought to be related to the study drugs. Information was recorded on antimalarial medication from public, private, or traditional sectors and on whether the woman’s family owned bednets. At about 34 weeks of gestation, women completed the antenatal component of the study with a blood test for haemoglobin concentration, malaria slide, and haemoglobin electrophoresis. HIV testing was done anonymously with linkage after deletion of personal identifiers. At the time of this blood test, all women who were anaemic were treated and followed up as appropriate, to reduce the likelihood of severe anaemia at the time of delivery. Women with haemoglobin below 10 g/dL were given folic acid and ferrous sulphate. Women with haemoglobin below 7 g/dL or with parasitaemia were given sulphadoxinepyrimethamine and those with haemoglobin below 6 g/dL were admitted to hospital. If study nurses suspected severe anaemia before 34 weeks’ gestation, haemoglobin was measured immediately. If the concentration was less than 7 g/dL, other investigations were done and treatment given, as described for the 34-week visit.
THE LANCET • Vol 353 • February 20, 1999
Full blood counts were done on a model M530 Coulter Counter. Haemoglobinopathies were identified by haemoglobin electrophoresis. Thick and thin blood films were prepared by standard giemsa staining, with counting of parasites per 200 white cells and calculation of counts per μL from the white-cell count. 100 high-power fields were read to confirm a negative film. Evidence of HIV infection was sought by GACPAT (Public Health Laboratory Service, London, UK).22 For women who gave birth in hospital, thick blood smears were taken at the time of delivery from the maternal surface of the placenta and examined for parasites and pigment by standard giemsa staining. Placental biopsy samples were prepared and examined by two people as described previously.23,24 Women were seen at least 4 weeks after delivery, so that stillbirths, neonatal deaths, maternal deaths, and maternal and neonatal morbidity could be ascertained.
Statistics Data entry, validation, and cleaning were done with DBase IV. Statistical analysis was carried out with STATA (release 4·0) by intention to treat for women who had a blood test at 34 weeks. The primary outcome was haemoglobin concentration below 8 g/dL at 34 weeks of gestation. Protective efficacy of sulphadoxine-pyrimethamine compared with placebo for specific endpoints was calculated from the risk ratios for those endpoints: protective efficacy=1003(1–risk ratio). Subgroup analyses were done on number of doses allocated and on areas of residence. For each potential confounder (educational status of woman and her partner, ownership of radio, type of house, marital status, number of other wives, sickle-cell trait, HIV infection, antimalarials received outside of the study protocol, area of residence, and bednet ownership), an adjusted risk ratio estimated by the Mantel-Haenszel technique was compared with the unadjusted risk ratio. For all such analyses, crude and adjusted risk ratios showed no evidence of confounding, hence the risk ratios presented are unadjusted. The sample size was based on that required to measure a 50% reduction in the proportion of women with haemoglobin below 8 g/dL with 80% power at 5% significance, separately for the population in the bednet-study area and the population not in that study area. We aimed to recruit 500 women from each area. However, a higher proportion of women than expected were not eligible because they were more than 30 weeks pregnant at first attendance, so we recruited only 429 women from the population from the bednet-study area, of whom 70% still used an insecticide-treated bednet. 1601 women screened 337 not eligible 268 ⬎30 weeks' gestation 46 severe anaemia or clinical malaria 11 severe pre-eclampsia or fetal death 10 refused 2 possible reaction to sulpha drugs 1264 women randomised
624 assigned placebo
59 had no blood test during third trimester 565 had blood test during third trimester
640 assigned sulphadoxine-pyrimethamine 73 had no blood test during third trimester 567 had blood test during third trimester
Trial profile
633
ARTICLES
Sulphadoxinepyrimethamine (n=640)
Placebo (n=623)*
Demography Ethnic group–Mijikenda Married
565 (88·3%) 538 (84·1%)
560 (89·9%) 506 (81·2%)
Educational status Able to read a letter No education Incomplete primary Primary complete Secondary complete
383 (59·8%) 206 (32·2%) 248 (38·8%) 111 (17·3%) 75 (11·7%)
370 (59·4%) 211 (33·9%) 243 (39·0%) 109 (17·5%) 60 (9·6%)
Educational status of husband/partner† Able to read Primary complete Secondary complete
557 (93·5%) 422 (70·8%) 161 (27·0%)
541 (95·3%) 417 (73·4%) 158 (27·8%)
358 (55·9%) 341 (53·3%) 144 (22·5%)
336 (53·9%) 334 (53·6%) 128 (20·6%)
81 (14·5%) 29 (5·1%) 23·0
80 (14·3%) 33 (5·8%) 23·1
Socioeconomic status Owns radio Latrine in compound House walls made of coral/stone (ie, not mud) Medical Sickle-cell trait‡ HIV positive§ Mean duration of gestation at recruitment (weeks)
Medication taken during pregnancy, before enrolment Sulphadoxine-pyrimethamine 18 (2·8%) Chloroquine treatment 36 (5·6%) Chloroquine prophylaxis 27 (4·2%)
12 (1·9%) 38 (6·1%) 25 (4·0%)
*Information available for only 623 of 624 women recruited. †n=596 for sulphadoxine-pyrimethamine group; n=568 for placebo group. ‡n=557 for sulphadoxine-pyrimethamine group; n=561 for placebo group. §n=571 for sulphadoxine-pyrimethamine group; n=565 for placebo group.
Table 1: Characteristics of intervention and control groups
Results Study population 1601 primigravid women attended clinic for the first time during the recruitment period (figure). 337 were not eligible. Of the 1264 women assigned sulphadoxinepyrimethamine or placebo, 233 (18%) were allocated three doses, 811 (64%) two doses, and 220 (17%) one dose. The study groups were similar in terms of ethnic group, education, socioeconomic status, gestation at recruitment, HIV status, frequency of sickle-cell trait, and drug history during the pregnancy (table 1). Efficacy Of 1264 women with a singleton pregnancy enrolled within the trial, 1132 (90%) had a study blood test: 1054 after 32 weeks’ gestation, 66 at 28–32 weeks’ gestation, and 12 at less than 28 weeks’ gestation. There was no
All participants
Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
82/567 (14·5%)
134/565 (23·7%)
39 (22 to 52)
<0·0001
By doses allocated One 12/99 (12%) Two 50/350 (14%) Three 20/118 (17%) Residence Rural south Rural north Total With ITBN No ITBN Urban
24/99 (24%) 88/378 (23%) 22/88 (25%)
50 (6 to 73) 39 (16 to 55) 32 (216 to 60)
0·027 0·002 0·16
36/198 (18%)
61/211 (29%)
37 (10 to 56)
0·011
26/243 (11%) 12/146 (8%) 14/97 (14%) 20/126 (16%)
48/236 (20%) 27/148 (18%) 21/88 (24%) 25/117 (21%)
47 (18 to 66) 55 (15 to 76) 40 (211 to 67) 26 (226 to 56)
0·004 0·011 0·102 0·271
difference in duration of gestation at blood test between the groups (p=0·44), and the reasons for early blood testing were similar in both groups. 11 women from the sulphadoxine-pyrimethamine group and 17 from the placebo group had the study blood test early owing to severe anaemia or clinical malaria. The most common reason for not having a study blood test was delivery before 35 weeks’ gestation (86 [65%] of 132). 118 (9·3%) women (59 from each group) did not take the number of doses that they had been allocated. There were no substantial differences between the groups in the reasons for non-completion of the assigned course, and no serious adverse reactions to the medication. Four women in the sulphadoxine-pyrimethamine group and three in the placebo group had study medication suspended because of suspected minor drug reaction. Among the 1132 women who had the third-trimester study blood test, a significantly higher proportion of the placebo group than of the sulphadoxine-pyrimethamine group were anaemic (haemoglobin ∂11 g/dL: 460 [81%] vs 431 [76%], p=0·026). Presumptive treatment with sulphadoxine-pyrimethamine also led to a lower rate of severe anaemia (haemoglobin <8 g/dL) with an overall protective efficacy of 39% (95% CI 22–52) and an absolute risk difference of 9·2% (p<0·0001, table 2). Efficacy was high and was significant even if only one dose of medication was given (p=0·027). The overall effect was similar in both rural populations, including the women who still had insecticide-treated bednets, though urban women did not seem to derive as much benefit from the intervention as rural women did (table 2). In some of the subsets, however, the sample sizes were small, so the statistical power was low. We also examined the effect of the intervention by place of recruitment (Vipingo or Kilifi District Hospital). There was less anaemia with sulphadoxinepyrimethamine than with placebo in all subgroups except women who gave an address from Kilifi south, but who attended the clinic at Kilifi District Hospital. In that subgroup, 16 (22%) of 73 women in the sulphadoxinepyrimethamine group and eight (11%) of 71 in the placebo group had severe anaemia (p=0·07). By contrast, the effect of the intervention in the rural population of Kilifi south who went to Vipingo clinic was large: 20 (16%) of 125 in the sulphadoxine-pyrimethamine group versus 53 (39%) of 137 in the placebo group had severe anaemia, giving a protective efficacy of 59% (p<0·0001). Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
All participants
30/567 (5·3%)
199/564 (35·3%)
85 (78–90)
<0·0001
By doses allocated One Two Three
4/99 (4%) 19/350 (5%) 7/118 (6%)
27/99 (27%) 134/377 (35%) 38/88 (43%)
85 (59–95) 85 (76–90) 86 (71–94)
<0·0001 <0·0001 <0·0001
21/198 (11%)
100/210 (48%)
78 (66–85)
<0·0001
4/243 (2%) 3/146 (2%) 1/97 (1%) 5/126 (4%)
66/236 (28%) 36/148 (24%) 30/88 (34%) 33/117 (28%)
94 (84–98) 92 (73–97) 97 (78–99) 86 (65–94)
<0·0001 <0·0001 <0·0001 <0·0001
Residence Rural south Rural north Total With ITBN No ITBN Urban
For placebo group; information on residence available for only 564 of 565 with antenatal haemoglobin measurements. ITBN=insecticide-treated bednet.
For placebo group; information on residence available for only 563 of 564 with antenatal blood film. ITBN=insecticide-treated bednet.
Table 2: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on proportion of women with severe anaemia
Table 3: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on proportion of women with peripheral parasitaemia
634
THE LANCET • Vol 353 • February 20, 1999
ARTICLES
Miscarriages* Stillbirths Perinatal deaths Neonatal deaths
Sulphadoxinepyrimethamine
Placebo
Protective efficacy (%) (95% CI)
p
4/630 (0·6%) 24/626 (3·8%) 39/626 (6·2%) 19/602 (3·2%)
4/615 (0·7%) 26/611 (4·3%) 49/611 (8·0%) 30/585 (5·1%)
2·7 (2287 to 76) 0·4 (22 to 3) 22 (217 to 48) 38 (28 to 65)
0·97 0·71 0·22 0·09
*Delivery at less than 24 weeks’ gestation.
Table 4: Effect of intermittent treatment doses of sulphadoxine-pyrimethamine compared with placebo on postnatal outcome
Mean haemoglobin concentration was significantly higher in the sulphadoxine-pyrimethamine group than in the placebo group (mean 9·7 g/dL [SD 1·76; range 2·7–14·6] vs 9·3 g/dL [1·85; 3·4–14·5], p<0·0001). 1131 women had peripheral blood film results. Among these women, malaria parasitaemia was less common in the sulphadoxine-pyrimethamine group than in the placebo group, with a protective efficacy of 85% (95% CI 78–90, p<0·0001); table 3). The effect was similar in all subgroups. There was little difference between the groups in reported treatment-seeking behaviour or medication during the study. Six women from the sulphadoxinepyrimethamine group and eight from the placebo group reported taking sulphadoxine-pyrimethamine, outside the study protocol, and 69 and 61 women, respectively, reported taking chloroquine. Five of 29 HIV-infected women in the sulphadoxinepyrimethamine group compared with nine of 33 from the placebo group had haemoglobin less than 8 g/dL, and three and ten, respectively, had peripheral parasitaemia. However, the differences were not significant. Among the women who gave birth in hospital, 16 (7·8%) of 205 in the sulphadoxine-pyrimethamine group and 29 (14·8%) of 196 in the placebo group had a placental smear positive for malaria (protective efficacy 47% [6–70], p=0·27); in 93 (45·4%) of 205 and 69 (34·2%) of 202, respectively, the placenta was negative for malaria on histology (protective efficacy 17% [3–29], p=0·021).
Postnatal follow-up Information on birth outcome was obtained from 1245 (98·5%) women. There were fewer stillbirths and neonatal deaths in the sulphadoxine-pyrimethamine group than in the placebo group, consistent with a protective efficacy of 22% for perinatal deaths and 38% for neonatal mortality, though these differences were not significant (table 4). There was no difference in the proportion of women delivering prematurely or in the mean duration of gestation at delivery between the groups. Of five babies who died and were reported to have been jaundiced, three were born to mothers who had not received sulphadoxine-pyrimethamine during pregnancy and two were born to mothers who had received the drugs more than 6 weeks before delivery. Five women died, four in the placebo group (one from chorioamnionitis at 30 weeks’ gestation, one with a history of acute abdominal pain antenatally, and two as a result of obstructed labour) and one in the sulphadoxinepyrimethamine group (with a history of severe pallor and acute diarrhoea at 30 weeks’ gestation). Of these women, the four who were tested for HIV were negative.
Discussion In this double-blind randomised controlled trial, presumptive treatment with sulphadoxine-pyrimethamine
THE LANCET • Vol 353 • February 20, 1999
greatly reduced the frequency of severe anaemia and was acceptable to women. Since the majority of primigravid women in Kenya attend for antenatal care (estimates are 95% for women attending antenatal clinics at least once in Kilifi District)25 and sulphadoxine-pyrimethamine is cheap (each dose costs US$0·17), this regimen should be practicable. The finding of a significant benefit from the intervention in women who attended late in pregnancy and so were assigned only one dose (table 2), suggests that even these women can benefit from the intervention and that health-workers should try as hard to give late attenders a single dose as to provide early attenders with two or three doses. However, the findings cannot be interpreted as meaning that one dose is sufficient treatment for all women. Women may develop severe lifethreatening anaemia secondary to malaria early in the second trimester, so presumptive treatment needs to be started as early as possible after 16 weeks of gestation. Furthermore, there may be differences between women who book for antenatal clinic late and women coming earlier in pregnancy, for example, in the proportion with illness. This regimen has the advantage over chloroquine that it is given to women when they attend antenatal clinics, thus helping to avoid problems of compliance. We did not detect any serious side-effects in mothers or infants, and there was no difference in minor adverse events between the groups. There is a small risk of severe adverse reactions to sulpha drugs,26 and a theoretical risk of kernicterus if these drugs are used within a short period before delivery.27 However, in an area with endemic malarial transmission, and much severe anaemia attributable to malaria in pregnancy, the benefits of this regimen outweigh these potential risks. No increased risk of kernicterus has been seen in this or other studies of sulphadoxine-pyrimethamine use in pregnancy.18,28 In our study, however, only 42 women received the drugs less than 14 days before delivery; the mean time between the last dose and delivery was 56 days. Although these results are reassuring, surveillance systems remain important for safety monitoring, if this regimen becomes routine. The rural population from the north had less anaemia than the southern population (table 2). These are geographically distinct populations, so there could be several reasons for this difference, including intensity of malaria transmission,21 variations in nutritional status, and variations in hookworm prevalence. The absolute difference in rates of anaemia between intervention and placebo groups was similar in all the rural subpopulations, including the population with insecticidetreated bednets. This finding confirms previous reports that bednets alone do not prevent the adverse outcomes of malaria in pregnancy.7 It also suggests that treatment with sulphadoxine-pyrimethamine is effective in a range of transmission settings, though was not as effective in the urban populations as in the rural populations. The only group of women who appeared to derive no benefit from the intervention were women recruited at Kilifi District Hospital who gave an address in Kilifi south. The proportion with anaemia among these women was similar to that in the urban population; these women may have been residing temporarily in the urban areas of Kilifi north. Severe anaemia in pregnancy has been reported as the main cause of up to 20% of maternal deaths in some hospital series in sub-Saharan Africa29,30 and 11–13% in community-based studies.31,32 In Kilifi District Hospital, 635
ARTICLES
severe anaemia was judged to be the primary factor in ten of 43 deaths over a 3-year period (EKD, CES; unpublished). Severe anaemia is also associated with substantial maternal morbidity. Malaria infection and severe maternal anaemia are also risk factors for low birthweight,3 one of the main risk factors for infant mortality.33 On the assumption that without treatment, anaemia at 34 weeks’ gestation will not resolve spontaneously, a decrease in the rate of severe anaemia in the third trimester is likely to have a great public-health impact. Malaria is an under-recognised cause of severe anaemia in areas of endemic transmission, since it is generally asymptomatic and not associated with peripheral parasitaemia. The delay in the development of an effective regimen to prevent malaria in pregnant women since the emergence of widespread chloroquine resistance has resulted in much morbidity and mortality. Unlike previous policies for pregnant women, this regimen does not involve continuous prophylaxis and is therefore unlikely to contribute significantly to the emergence of drug resistance. Use in pregnant women will form only a small part of total use. Nonetheless, the likely emergence of resistance to sulphadoxinepyrimethamine is a concern, and research to identify ways of preventing such resistance and of finding alternative treatments should continue in parallel with implementation of policy. Contributors Caroline Shulman designed the study, coordinated and supervised data collection, did the analysis, and wrote the paper. Ed Dorman carried out all ultrasound scans, supervised data collection, and was involved in the clinical management. Felicity Cutts contributed to the study and advised on the analysis and technical and practical issues. Ken Kawuondo undertook and supervised day-to-day laboratory work. Judith Bulmer undertook the histological examinations of the placenta. Norbert Peshu contributed to the study design and advised on practical issues. Kevin Marsh contributed to the study design and advised on the analysis and technical and practical issues. All investigators contributed to the writing of the paper.
Acknowledgments This paper is published with the permission of the Director of KEMRI. The work was supported by the UK Department for International Development and KEMRI. We thank Peter Smith, Brian Greenwood, Bernard Brabin, and Cathy Waruiru, who were members of the safety monitoring committee; Bob Snow for his advice, all at the Kilifi research unit, particularly Brett Lowe who was responsible for managing the laboratory work; midwives Jane Mwendwa, Judith Peshu, and Ann Muhoro; the study fieldworkers; Oona Campbell and Tom Marshall for their advice and valuable comments on the paper; and John Parry (Hepatitis and Retrovirus Laboratory, Public Health Laboratory Service, London) who assisted in the supply of kits and advice on HIV testing. KM is a Wellcome Trust senior research fellow (631342).
8
9
10
11 12
13
14
15
16
17
18
19
20
21
22
23
24
25
References 1
2
3 4
5
6
7
Shulman C, Graham W, Jilo H, et al. Malaria as an important cause of anaemia in primigravidae: evidence from a district hospital in coastal Kenya. Trans R Soc Trop Med Hyg 1996; 90: 535–39. Steketee RW, Wirima JJ, Hightower AW, et al. The effect of malaria and malaria prevention in pregnancy on offspring birthweight, prematurity, and intrauterine growth retardation in rural Malawi. Am J Trop Med Hyg 1996; 55 (suppl 1): 33–44. Brabin B. The risks and severity of malaria in pregnant women. Appl Field Research in Malaria reports WHO 1991; no1. Steketee RW, Breman JG. Malaria infection in pregnant women in Zaire: the effects and potential for intervention. Ann Trop Med Parasitol 1988; 82: 113–20. McDermott JM, Heymann DL, Wirima JJ, et al. Efficacy of chemoprophylaxis in preventing Plasmodium falciparum parasitaemia and placental infection in pregnant women in Malawi. Trans R Soc Trop Med Hyg 1988; 82: 520–23. Descowitz RS, Alpers MP. Placental Plasmodium falciparum parasitaemia in East Sepik women of different parity. Ann Trop Med Parasitol 1992; 86: 95–102. Shulman C, Dorman E, Talisuna A, et al. A community randomised controlled trial of insecticide treated bednets for the prevention of
636
26
27
28 29 30 31
32
33
malaria and anaemia among primigravid women on the Kenyan Coast. Trop Med Int Health 1998; 3: 197–204. Mutabingwa TK, Malle LN, de Geus A, Oosting J. Malaria chemosuppression in pregnancy II: its effect on maternal haemoblobin levels, placental malaria and birth weight. Trop Geogr Med 1993; 45: 49–55. Greenwood BM, Greenwood AM, Snow RW, Byass P, Bennett S, Hatib-N’Jie AB. The effects of malaria chemoprophylaxis given by traditional birth attendants on the course and outcome of pregnancy. Trans R Soc Trop Med Hyg 1989; 83: 589–94. Fleming AF, Ghatoura GBS. The prevention of anaemia in pregnancy in primigravidae in the guinea savanna of Nigeria. Ann Trop Med Parasitol 1986; 80: 211–33. Gilles HM, Lawson JB, Sibelas M, Voller A, Allan N. Malaria, anaemia, and pregnancy. Ann Trop Med Parasitol 1969; 63: 245–63. Heyman DL, Steketee RW, Wirima JJ, McFarland DA, Khoromana CO, Campbell CC. Antenatal chloroquine prophylaxis in Malawi: chloroquine resistance, compliance, protective efficacy and cost. Trans R Soc Trop Med Hyg 1990; 84: 496–98. Gülmezoglu AM, Garner P. Malaria in pregnancy in endemic areas. (Cochrane Review) In: The Cochrane Library : Issue 4, 1998. Oxford: Update Software. Dolan G, ter-Kuile FO, Jacoutot V, et al. Bed nets for the prevention of malaria and anaemia in pregnancy. Trans R Soc Trop Med Hyg 1993; 87: 620–26. D’Alessandro U, Langerock P, Bennett S, Francis N, Cham K, Greenwood BM. The impact of a national impregnated bed net programme on the outcome of pregnancy in primigravidae in The Gambia. Trans R Soc Trop Med Hyg 1996; 90: 487–92. Alonso PL, Lindsay SW, Armstrong JRM, et al. The effect of insecticide-treated bed nets on the mortality of Gambian children. Lancet 1991; 337: 1499–502. Nevill CG, Some ES, Mung’ala VO, et al. Insecticide treated bednets reduce mortality and severe morbidity from malaria among children on the Kenyan coast. Trop Med Int Health 1996; 1: 139–46. Schultz LJ, Steketee RW, Macheso A, Kazembe P, Chitsulo L, Wirima JJ. The efficacy of antimalarial regimens containing sulfadoxine-pyrimethamine and/or chloroquine in preventing peripheral and placental plasmodium falciparum infection among pregnant women in Malawi. Am J Trop Med Hyg 1994; 51: 515–22. Mbogo CNM, Snow RW, Khamala CPM, et al. Relationship between Plasmodium falciparum transmission by vector populations and the incidence of severe disease at nine sites on the Kenyan Coast. Am J Trop Med Hyg 1995; 52: 201–06. Snow RW, Schellenberg JR, Peshu N, et al. Periodicity and space-time clustering of severe childhood malaria on the coast of Kenya. Trans R Soc Trop Med Hyg 1993; 87: 386–90. Snow RW, Omumbo JA, Lowe B, et al. Relation between severe malaria morbidity in children and level of Plasmodium falciparum transmission in Africa. Lancet 1997; 349: 1650–54. Parry JV, Mortimer PP. An immunoglobulin G antibody capture particle-adherence test (GACPAT) for antibody to HIV-1 and HTLV1 that allows economical large-scale screening. AIDS 1989; 3: 173–76. Bulmer JN, Rasheed FN, Francis N, Morrison L, Greenwood BM. Placental malaria I: pathological classification. Histopathology 1993; 22: 211–18. Bulmer JN, Rasheed FN, Morrison L, Francis N, Greenwood BM. Placental malaria II: a semi-quantitative investigation of the pathological features. Histopathology 1993; 22: 219–25. Kirumbi LW, Wassuna A, Nabwani G, Mpanju-Shumbusho W. Prevalence of and factors associated with perinatal morbidity and mortality in Kirinyaga and Kilifi Districts in Kenya. Project report, 1996. Bjorkman A, Phillips-Howard PA. Adverse reactions to sulpha drugs: implications for malaria chemotherapy. Bull World Health Organ 1991; 69: 297–304. Walker PC. Neonatal bilirubin toxicity: a review of kernicterus and the implications of drug-induced bilirubin displacement. Clin Pharmacokinet 1987; 13: 26–50. Phillips-Howard PA, Wood D. The safety of antimalarial drugs in pregnancy. Drug Saf 1996; 14: 131–45. Armon PJ. Maternal deaths in Tanzania. Trans R Soc Trop Med Hyg 1979; 73: 284–88. Mtimavalye LAR, Lisasi D, Ntuyabaliwe WK. Maternal mortality in Dar es Salaam, Tanzania. East Afr Med J 1980; 57: 111–18. Boerma JT, Mati JKG. Identifying maternal mortality through networking: results from coastal Kenya. Stud Fam Plann 1989; 20: 245–53. MacLeod J, Rhode R. Retrospective follow-up of maternal deaths and their associated risk factors in a rural district of Tanzania. Trop Med Int Health 1998; 3: 130–37. McCormick MC. The contribution of low birth weight to infant mortality and childhood mortality. N Engl J Med 1985; 31: 82–90.
THE LANCET • Vol 353 • February 20, 1999