Study area and participants

A total of 1,084 pregnant women who were attending antenatal clinics (ANC) for the first time during current pregnancy in two hospitals in semi-urban areas of Lagos – Ajeromi General Hospital in Ajegunle (501) and St Kizito Primary Health Center in Lekki (583) – and who consented to the study were enrolled between March 2007 and February 2008. These antenatal clinics were chosen because they record high antenatal clinic attendance. In sub-Saharan Africa, high antenatal clinic visits are common [23]. The antenatal clinics met this criterion of providing four qualitative visits by pregnant women. Since adequate malaria control strategies were expected to be initiated after booking at antenatal clinics, the women were recruited on the first day of registration for ANC.

This study was approved by the Ethics and Experimental Committee of the College of Medicine of the University of Lagos Lagos, Nigeria. All participants gave informed consent after the purpose of the research was explained to them. Participation in this research was voluntary as participants could decline to participate or withdraw from participation even after giving consent at any stage of the research. Patients that declined to participate were not denied access to the available routine care.

Data collection and laboratory analysis

The study design was cross-sectional. The minimum study population was 384 as estimated with the Statcalc software of Epi Info 6 (Center for Disease Control, Atlanta) for Population Survey or Descriptive Study using random sampling. The sample size was calculated using the following assumptions: the population size of Lagos based on the national census was 9,013,534; the population of pregnant women in 2006 was 396,595 (4.4% of the Lagos population) (UNFPA http://www.unfpa.org/emergencies/manual/9a5.htm. Accessed 19th September 2006); the population of pregnant women in 2007 (growth of 2.34%) was 405,875; expected frequency of malaria in pregnancy of 10%; worst acceptable frequency of malaria in pregnancy of 7.0% and a confidence level of 95%. However, the total number of pregnant women recruited in the study was 1,084.

Demographic data, and information on history of fever during current pregnancy, treatment and preventive measures adopted were collected from the pregnant women using a semi-structured questionnaire which was administered by a trained interviewer. Peripheral blood by venepuncture was used to collect 1mL of blood for malaria microscopy and packed cell volume determination.

The total leukocyte count determination was done using the improved Neubauer Chamber, as described in detail by Baker and Silverton [24]. Briefly, 20 μL taken with an adjustable micropipette (P20 Pipetman, Gilson) of blood was mixed with 380 μL of Turk’s solution (2% Acetic acid tinged with gentian violet) taken with an adjustable micropipette (P1000 Pipetman, Gilson) to give a final 1:20 dilution. The red cells were lysed leaving the leukocytes. The leukocytes were counted using the New Improved Neubauer Chamber.

The haematocrit level was determined by filling a capillary tube (Hawksley, England) up to 75% with well-mixed anticoagulated blood, sealed at one end with Cristaseal, (Hawksley, England) and spun at 13,000 rpm for five minutes in a Hawksley haematocrit centrifuge (Hawksley, England). The percentage of the packed cells was read with a Hawksley Haematocrit Reader (Hawksley, England).

Malaria microscopy

Thin and thick blood films for the diagnosis of malaria were prepared on the same slide. For the thick film, 12 μL of blood was taken with an adjustable micropipette (P20 Pipetman, Gilson) and spread over a diameter of 15 mm, while 2 μL of blood was used to make the thin film. The slides were made in duplicates and labeled appropriately. The thin film end of the slide was fixed by dipping the prepared film in absolute methanol for one to two seconds, and both films were allowed to dry for 24–48 hours and subsequently stained with 3% Giemsa at pH 7.2 for 45 minutes.

The stained slides were read by two certified microscopists. Discrepant parasite detection and parasite count readings between the two microscopists were resolved by a re-reading of the slides or by employing a third microscopist before the final result was determined. The mean parasite counts of the two readers were accepted if the discrepancy of the two readings was less than 20%. A slide was confirmed negative when 100 high power fields have been examined under x100 oil immersion objective lens.

The absolute parasite density was calculated using the formula: [# of parasites counted x total leukocyte count] / # of leucocytes counted.

Data analysis

The data generated from the study were analysed using EPI INFO 2002 statistical software (CDC, Atlanta, USA). Tests for associations and differences were done by chi-square analysis, risk ratio analysis, Fisher Exact test, analysis of variance and a Kruskal-Wallis test where appropriate. Test of statistical significance was set at p value less than 0.05 at 95% confidence interval.

Relation to age and gravidity

In this study, maternal age was associated with malaria prevalence, showing that a pregnant woman of young maternal age is at the greatest risk of malaria infection, as well as having the highest parasite densities. Similar findings have been reported by other authors in Gabon and Eastern Sudan where malaria prevalence was observed to decrease as age increased [28, 29].

The effects of malaria in pregnancy have been noted to be lower in multigravidas than in other gravidities as a result of acquisition of specific immunity to placental malaria due to previous exposure [6, 14, 30]. Acquired specific immunity accumulates with subsequent infection and subsequent pregnancies [14]. In this study, gravidity was not associated with either malaria prevalence or the level of parasitaemia. Young maternal age has been reported as a more important risk factor than gravidity [31] and is consistent with this study. The level of acquired immunity, which is associated with the number of malaria infections during pregnancy, was not determined in this study. If this was done, it could have explained the lack of association between malaria prevalence and gravidity. In a malaria endemic area such as Lagos, it is possible that the women would have had a number of encounters with malaria infection prior to booking at the antenatal clinics used in this study. It could be that there was no difference in the level of specific immunity of the study participants based on gravidity. Moreover, some authors have also observed that women in their second pregnancies are almost as susceptible to malarial infections as those that are pregnant for the first time [7, 32]. Early attendance and participation in focused ANC services is recommended to all pregnant women especially the primigravidae so as to reduce the risk of malaria infection in pregnancy.

Relation to education and gestation age

Education and the gestation ages of the women at the time of registration for ANC at the clinics were not significantly associated with malaria infection. Low educational level was not associated with malaria infection. This may be attributed to massive radio and television campaigns on malaria prevention strategies and appropriate treatment options in Lagos. These campaigns are also regular features in child welfare clinics all over Lagos and are conducted in English, as well as local languages.

Relation to insecticide spray

This is the first study to report the effect of the use of personal/household insecticide spray on malaria in pregnancy in Nigeria. The use of insecticide spray was very common among pregnant women in this study. Also, the frequency of use of these insecticide sprays was high, with 67.2% of the women using insecticide spray more than once a week. This probably accounted for its significant impact on malaria infection.

Relation to insecticide treated nets (ITN)

Insecticide treated nets (ITN) coverage in this study population was far from the 60% target – of pregnant women sleeping under an ITN – that was set by African Heads of States at the Roll Back Malaria Summit held in Abuja in 2000 [33]. The higher prevalence of malaria in pregnant women that used an ITN as opposed to those that used any net suggested that strict adherence to sleeping under an ITN may be a challenge as there were complaints of excessive sweating (data not shown) caused by poor ventilation, which was further exacerbated by the frequent absence of public power supply running electric fans and air conditioners. Some other studies have also reported that net ownership does not necessarily translate to usage [34, 35].

Despite reports that the use of ITNs substantially reduced the risk of malaria in pregnancy [36], the use of bed nets, whether treated or not, did not have a significant impact on malaria infection in this study. Marchant and co-workers [33] also reported a non-significant but modest impact of ITN on malaria in pregnancy. The suggested reasons for the conflicting results on the efficacy of ITN are: a) compliance with sleeping under a bed net; and b) the complex vector populations with exophagic\exophilic and early biting behaviour. The high rate of the use of insecticide spray in this study may also be responsible for the very low rate of the use of ITNs.

Relation to the combination of preventive methods

The combined use of insecticide spray and ITN resulted in a reduction in malaria infection though this reduction was not significant. None of the women that used a combination of the three preventive methods had malaria infection. Education and age were significantly associated with the combined use of insecticide spray and ITN. This could be explained by the increased financial capacity to purchase insecticide sprays. The very strong association with age could be due to poor financial capacity of women in the 15–19 years of age group that may have resulted in the poor use of both strategies. Pregnant women above 24 years of age were likely to be employed thus resulting in a higher frequency of use (>80%) of this combination. Occupation was not significantly associated with the use of combination interventions. This could be the result of the challenge in grading the various professions into low- or high-income professions.

Relation to chemoprophylaxis

The use of un-recommended malaria chemoprophylactic medications such as chloroquine (CQ) did not significantly reduce malaria infection among the pregnant women. This finding was not surprising because high CQ failure rates had been reported in Southwest Nigeria. The various CQ failure rates reported ranged from 37.9% to 59.1% [37–39]. Coulibaly et al.[40] also reported a failure rate of 46.7% among pregnant women treated with CQ in Burkina Faso. The current Nigerian National guideline for malaria prevention and control during pregnancy replaced CQ and pyrimethamine with SP as malaria chemoprophylaxis during pregnancy because of issues of compliance with the dosing regimen and resistance [41]. However, the use of un-recommended antimalarial medications as chemoprophylaxis, is likely to increase the risk of pregnant women having a malaria infection.