Perennial transmission of malaria in the low altitude areas of Baringo County, Kenya

The study was conducted in three sub-counties (Marigat, Baringo North and Baringo Central) of Baringo County between February 2015 and February 2016. Baringo County is located within the Rift Valley and lies between longitude 35.602°–36.277°E, and latitude 0.541°–0.723°N at altitudes ranging between 870 and 2499 m above sea level (asl). The study area represents arid and semi-arid parts of Baringo County. The study area was divided into four strata running parallel to each other in a North–South direction. These strata were identified based on the hydrology, soil types, rainfall patterns, vegetation cover and altitude (Fig. 1). The four zones included; lowlands (900–1000 m asl), midlands (1000–1500 m asl), highlands (1500–2300 m asl) and riverine (1000–1200 m asl). The lowland zone lies to the east of the study area and receives mean annual rainfall of about 650 mm with temperatures ranging between 30 and 37 °C. The area is characterized by both lakes (Lake Baringo and Lake 94) and perennial rivers (Perkerra and Molo) with poor drainage soil types. The midland zone is characterized by well drained soil types and is interspersed with dry river beds that flow only after heavy rains. The highland zone receives an average rainfall that ranges between 1000 and 1500 mm per year with well drained soil types, while the riverine zone borders the Kerio Valley to the west of the study area. Lake Kamnarok is the major water body within this zone where the soil is poorly drained and prone to flooding.

According to 2009 Kenya national population census, Baringo County has a population of 555,561 people distributed in 11,075.3 km². The temperatures range from a minimum of 10 °C to a maximum of 37 °C in different parts of the county. The region has two distinct weather patterns with the cold months (June and July) and the hot months (January and February). It experiences two rainy seasons, March to June (long rains) and November (short rains). It receives between 1000 and 1500 mm of rainfall annually in the highlands and 600 mm in the lowlands. Malaria transmission in this region is seasonal with peaks during the rainy season. The current incidence of malaria is about 12% based on outpatient visits to local health facilities [12]).

The study was longitudinal in design where 15 public day primary schools were selected from the four zones based on their close proximity to mosquito breeding (aquatic) habitats. These breeding sites had been identified by a parallel entomological study by the same team, and were equitably spatially distributed across the study area. From these, the school closest to a breeding site included in the study was conveniently selected, and based on the populations of the respective schools, 15 were adequate for the desired sample. From these schools, at least 1540 pupils aged between 5 and 15 years were required for the study, was estimated using a formula proposed by Cochran [13]. The study, however, enrolled 1668 pupils.where: n is the sample size, Z is the value for the selected alpha level, e.g. 1.96 for (0.25 in each tail) a 95 percent confidence level. p is the estimated proportion of an attribute that is present in the population (0.5). q is 1 − p. (p)(q) are the estimate of variance. e is the acceptable margin of error for proportion being estimated (0.05).

Following parental or guardian informed consent, 1668 pupils were recruited to mitigate against possible loss during follow-ups (see Table 1). During baseline study (January/February 2015) and each of the follow-ups (June/July 2015, October/November 2015, and January/February 2016), CareStart™ HRP-2(Pf) (Cat No. G0140 RDT Access Bio) kits were used to test for the presence of parasite antigen in whole blood, according to the manufacturers’ instructions. Pupils were physically examined and vital signs (weight, body temperature general health) recorded. Pupils were also asked for history of medication for febrile illnesses in the previous 4 weeks, and only those without were enrolled. One finger prick blood sample (20 µl) was then collected for smears and RDT. Both thick and thin smears were prepared in the field on the same slide and stored for subsequent laboratory examination. In the laboratory, thin and thick blood films were stained using 10% Giemsa stain. They were then examined under (100×) oil immersion lens. Slide examination was carried out by two microscopists. For positive slides, number of parasites were counted against 200 leucocytes and quantified as parasite density/µl of blood. Slides were considered negative when no parasite was detected only after examining 100 microscopic fields… All individuals diagnosed positive for Plasmodium spp. were treated for malaria using the recommended drug (Artemether + Lumefantrine) by the nurse; the pupils were given the first dose instantly, and instructed to complete the remaining doses at home. The average monthly rainfall data was downloaded from IRI/LDEO climate data library from January 2015 to January 2016 [14].

Incidence was estimated as the number of cases per 1000 person-months at risk. When detecting cases, any pupil who tested positive for malaria or missed on the testing was considered out of the study and was not followed. Time at risk for each pupil was then calculated based on how long the pupil was followed during the entire study period. Those who tested positive at baseline survey were not included when calculating time at risk since they all had time zero at risk. The incidence rate was calculated at an interval of 4 months for a period of 8 months. Those who tested positive during the first follow up had 2 months time at risk each, while in the second follow up, each had 6 months at risk. Those who were followed throughout the study period each had 8 months time at risk. Date of testing during baseline and subsequent follow up was recorded to determine time interval between each surveys.

Data obtained was entered in MS Excel spreadsheet, cleaned and malaria incidence was analysed using STATA version 12 (Stata Corporation, College station, Texas, USA). R version 3.0.3 was used to calculate point prevalence at 95% confidence interval of proportions while frequencies and cross tabulation to determine asymptomatic cases were calculated using SPSS version 22.

Ethical approval to conduct the study was obtained from Kenyatta National Hospital and University of Nairobi Ethics and Research Committee (P70/02/2013). Authority to conduct research was granted by the Director of Health Services, Baringo County. Written informed consent was obtained from each parent or guardian, and only willing children were enrolled into the study.

A total of 1668 primary school pupils were recruited and tested for malaria parasites during the baseline survey carried out in the dry season (January/February 2015) (Table 2). Of the total number tested, 175 (10.5%) (95% CI 9.1–12.1) pupils were positive for P. falciparum infections. The highest point prevalence was recorded from the riverine zone (22.8%) (95% CI 19.9–26.0) followed by the Lowlands (0.2%) (95% CI 0.01–1.5). Both midland and highland zones had no cases of Plasmodium species infections (Fisher’s exact test = 0.005).

During the first follow-up conducted towards the end of long rains (June/July 2015) the overall prevalence dropped from 10.5% (95% CI 9.1–12.1) to 2.6% (95% CI 1.8–3.6), although this difference was not statistically significant (Fisher’s exact test = 1.000). The number of pupils who were tested also dropped from 1668 to 1372 (17.7% loss to follow up). All positive cases were from the riverine zone giving a prevalence of 6.5%.

The second follow up was carried out during the short rains (October/November 2015). Malaria prevalence rose up to 5.5% (95% CI 4.3–6.9) from 2.6% (95% CI 1.8–3.6) (Fisher’s exact test = 0.551). The loss to follow up rose to 23.4%. The prevalence of malaria in the riverine zone was 11.4% (95% CI 9.1–14.4), while that in the lowland Zone was 0.6% (95% CI 0.1–2.5). No positive cases were reported in the highland and midland zones. The third follow-up was conducted immediately after the end of the El Niño rains (January/February 2016). The riverine zone recorded the highest prevalence of malaria infection (12.7%) (95% CI 5.0–7.7) compared to highlands (1.9%) (95% CI 0.7–4.7) and the Lowlands with 0.3% (95% CI 0.02–2.0). The overall point prevalence of malaria infection was 6.2% (95% CI 5.0–7.7). The midland Zone had no malaria cases during the entire period of the study.

In general, prevalence varied with altitude where the riverine zone recorded higher prevalence followed by the lowland and the highland (Fisher’s exact test = 0.005) (Table 1). Plasmodium species infection was slightly high during dry season compared to long rain season and short rain season.

The overall malaria incidence within the entire study area was estimated at 6/1000 person-months. This indicated that at least six people out of 1000 were at risk of P. falciparum infection every month. Malaria incidence varied greatly when categorized into geographical zones, with the riverine recording the highest (RR = 40.2 (95% CI 7–1623). While the highland and midland zones recorded zero incidence during the entire study period, the lowland had an incidence of 0.5/1000 person-month. The riverine zone had ant incidence of 14/1000 person-months (Table 3).

The risk of Plasmodium species infection was further analysed by age groups; 5–9 and 10–15 years old. Results indicated that those in the 10–15 years category were at a greater risk of P. falciparum infection (7/1000 person-months) when compared to 5–9 years age group (4/1000 person-months) although this was not statistically significant (RR = 1.63 (95% CI 0.9–3.1). Similarly, the rate of Plasmodium species infection between males and females was not significantly different within the study area. IRR = 1.02 (95% CI 0.55–1.88) (Table 4).

Asymptomatic Plasmodium species infection (infected individuals with no clinical signs or symptoms) was determined using both RDT and microscopy. During baseline survey, 61.1% (95% CI 0.53–0.68) of the positive cases by RDT were asymptomatic. The proportion of asymptomatic cases were relatively high in both first follow up and second follow up, 65.7% (95% CI 0.48–0.8) and 62.9% (95% CI 0.5–0.7), respectively. During the third follow up, 48.9% of the positive cases were asymptomatic. Pearson Chi square was used to determine the possible association between asymptomatic Plasmodium species infection to gender and age groups. Males were slightly more asymptomatic than females, however, the results indicated no statistically significant difference (Pearson χ² = 2.8885, df = 3, p = 0.409). Similarly, there was no difference of asymptomatic Plasmodium species infection between children aged 5–9 and 10–15 years (Pearson χ² = 0.6746, df = 3, p = 0.879) (Table 5).

Baringo County has a bimodal rainfall pattern with the long rains falling between April and July, and the short rains between August and November. The dry season occurs between December and March. Malaria transmission was highest during the months of January/February followed by October/November and lowest during June and July (Fig. 2).