Impact of Agricultural Irrigation on Anemia in Western Kenya

ABSTRACT.

Expanding agricultural irrigation efforts to enhance food security and socioeconomic development in sub-Saharan Africa may affect malaria transmission and socioeconomic variables that increase the risk of anemia in local communities. We compared the prevalence of anemia, Plasmodium falciparum infection, and indicators of socioeconomic status related to nutrition in communities in Homa Bay County, Kenya, where an agricultural irrigation scheme has been implemented, to that in nearby communities where there is no agricultural irrigation. Cross-sectional surveys conducted showed that anemia prevalence defined by WHO criteria (hemoglobin < 11 g/dL) was less in communities in the irrigated areas than in the non-irrigated areas during the wet season (38.9% and 51.5%, χ² = 4.29, P = 0.001) and the dry season (25.2% and 34.1%, χ² = 7.33, P = 0.007). In contrast, Plasmodium falciparum infection prevalence was greater during the wet season in irrigated areas than in non-irrigated areas (15.3% versus 7.8%, χ² = 8.7, P = 0.003). There was, however, no difference during the dry season (infection prevalence, < 1.8%). Indicators of nutritional status pertinent to anemia pathogenesis such as weekly consumption of non–heme- and heme-containing foods and household income were greater in communities located within the irrigation scheme versus those outside the irrigation scheme (P < 0.0001). These data indicate that current agricultural irrigation schemes in malaria-endemic communities in this area have reduced the risk of anemia. Future studies should include diagnostic tests of iron deficiency, parasitic worm infections, and genetic hemoglobin disorders to inform public health interventions aimed at reducing community anemia burden.

INTRODUCTION

Twenty-five to 30% of the global population is estimated to have anemia based on a lower blood hemoglobin (Hb) concentration that is less than that of healthy age- and gender-matched residents of high-income countries in non-tropical areas of the world.^1,2 Women of childbearing age and children in low- and middle-income countries of Africa are particularly vulnerable to anemia.³ The underlying causes of anemia include nutritional deficiencies of essential elements needed for Hb synthesis (e.g., iron, and vitamins A and B [riboflavin, B12, folate]), malaria and parasitic worm infections (e.g., hookworm and schistosomiasis), and genetic disorders of Hb synthesis (e.g., sickle cell).^4–7 Severe anemia during childhood, defined as a blood Hb concentration < 7 g/dL, is associated with slow cognitive development and impaired physical growth.^8–10

Socioeconomic and demographic factors (e.g., household income, education, gender, and pregnancy) also play key roles in achieving proper nutrition, and thus affect the risk of anemia.¹¹ Studies performed in western Kenya during the first 10 years of the 21st century found that poor nutrition was a primary cause of anemia in children.^12–14 Homa Bay County, located near the Lake Victoria basin, has historically experienced perennial malaria transmission and frequent food shortages resulting from prolonged droughts.^15,16 To alleviate this food insecurity, the Kenyan government supported construction of a concrete channel-based irrigation scheme, the Kimira-Oluch Small Holder Farm Improvement Project. Since completion of the irrigation scheme in 2016, a variety of crops (rice, sorghum, millet, beans, and maize) and vegetables (kale and spinach) has been grown. Increased availability of these crops for local consumption as well as a source of income resulting from sales to residents of communities outside the irrigation scheme appear to have improved the nutritional and socioeconomic status of local communities. The objective of this study was to evaluate the degree of improvement with respect to the prevalence of anemia and Plasmodium falciparum infection and socioeconomic status. Accordingly, we compared these variables in residents of communities within the irrigation scheme to nearby communities outside the irrigation scheme.

MATERIALS AND METHODS

Study area.

We conducted a study in the Rangwe and Rachuonyo South sub-counties of Homa Bay County in 2018 and 2019, and targeted communities located near canals within irrigated and non-irrigated areas located 5 to 10 km from irrigation canals (Figure 1). Two annual rainy seasons are associated with increased malaria transmission in Homa Bay County: the “long rains” in April through June and “short rains” in September through November. Annual rainfall ranges from 250 to 1,000 mm. The Kenyan government completed the Kimira-Oluch Smallholder Farm Improvement Project, a gravity-fed irrigation system with a concrete canal system, in 2016.¹⁵ Canal maintenance has been challenging, resulting in leaks and overflow during water release. Subsistence farming is the primary economic activity in local communities. There are two private and five public health-care facilities in the irrigated areas, and one private and five public health-care facilities in the non-irrigated areas. Homa Bay County is a malaria-endemic area, where the prevalence of P. falciparum blood-stage infection has historically been ≥ 20%.¹⁶ Annual indoor residual spraying with Actellic insecticide commenced in 2018. The prevalence of malaria infection is thought to have decreased substantially since then (Omondi CJ et al, unpublished data).

Figure 1.

Map of study area (part of Rangwe and Rachuonyo South sub-counties) in Homa Bay county (see Inset in the map of Kenya). This figure appears in color at www.ajtmh.org.

Study design and laboratory methods.

Cross-sectional surveys of community residents 6 months and older in the irrigated and non-irrigated areas were conducted during the June 2018 wet season and the February 2019 dry season. Blood samples were collected by fingerpick to determine Hb levels and P. falciparum blood-stage infection. To measure Hb concentration, fresh blood was placed in the optical window of a micro-cuvette HemoCue HB 201+ analyzer (Angelholm, Sweden). Based on WHO-recommended criteria,¹⁷ Hb concentration of each study participants was classified as having severe, moderate, mild, or no anemia, after adjusting for age and gender, as shown in Table 1.

Table 1.

Hemoglobin values for classification of anemia

Study participants	Hemoglobin concentration, g/dL			Non-anemic, g/dL
	Severe	Moderate	Mild
Children 6–59 months	< 7	7–9.9	10–10.9	≥ 11
Children 5–11 years	< 8	8–10.9	11–11.4	≥ 11.5
Children 12–14 years	< 8	8–10.9	11–11.9	≥ 12
Non-pregnant women ≥ 15 years	< 8	8–10.9	11–11.9	≥ 12
Men ≥ 15 years	< 8	8–10.9	11–12.9	≥ 13

Microscopic inspection of blood smears was performed to determine Plasmodium infection status and parasite density. Thick and thin blood smear slides were stained with 10% Giemsa for 15 minutes, washed, air-dried, and examined by microscopy. A minimum of microscopic fields containing 200 leukocytes were examined at ×100 magnification. Two independent microscopists read blood smears for quality control. A third reading was performed by a different microscopist if there was a discrepancy. Parasite density was determined by counting the number of parasite-infected red blood cells in microscopic fields that included 200 leukocytes and multiplying by 40, with the assumption that there were 8,000 leukocytes/µL blood for all study participants.

With respect to socioeconomic status and intake of nutrients pertinent to Hb production, the head of each household was interviewed using a pretested questionnaire to determine the source of income, annual household income, and consumption of heme- and non–heme-containing foods, and vitamins A and B in the previous week. Food intake was quantified according to one of four categories: no intake, intake 1 to 3 days, intake 4 to 6 days, and daily intake for the past week. Annual household income was categorized as severe poverty (annual income less than U.S. dollars [USD] 365), moderate poverty (income USD 365–1,095), and non-poor (income > USD 1,095).

Selection of study communities.

As described by Galway et al.,¹⁸ we adopted a two-stage cluster-sampling method, with clusters (defined as a collection of households with a single geographic unit) and households serving as the primary and secondary sampling units, respectively. The size (radius) of each cluster ranged from 250 to 500 m, with a population size of ∼200 to 250 people. Briefly, 20 clusters were selected based on population size. Ten clusters were located in irrigated areas and 10 in the non-irrigated areas. Six clusters were selected at random from both regions. After a cluster was identified, the total number of households within each cluster was established and assigned unique numbers. Based on the required sample size, 20 households per cluster were selected randomly from the previously marked households. Every household member (i.e., a permanent resident age ≥ 6 months), was enrolled in the study after obtaining signed informed consent or an assent form from children 5 to < 18 years. Pregnant women were excluded from the study. However, pregnant women were tested for anemia and malaria infection, and were referred to the local health-care facility for further assessment and treatment. The total number of participants who met the inclusion criteria was 836. The sample size was calculated using the formula: n = Z² P(1 – P)/D², where D represents precision (0.05), P is proportion (0.27), Z is the CI (1.96), and n is the required minimum sample size. Because anemia determination was a major objective of the study, an estimated malaria prevalence of 27% was assumed in the sample size calculation. The adopted cluster sampling procedure, hence, the sample size, was multiplied by a design effect of two.¹⁹ The required sample size was 606 people. To avoid attrition during surveys, we decided to enroll all eligible participants (N = 836). There were 712 and 826 participants in the 2018 wet season and 2019 dry season surveys, respectively.

Statistical analysis.

Data were entered into Excel spreadsheets and analyzed using GraphPad Prism version 8 and SPSS version 20 (SPSS Inc., Chicago, IL). The χ² or Fisher’s exact tests were used to analyze categorical variables. Means (Hb concentration, malaria parasite densities) between groups residing in irrigated and non-irrigated areas were tested for significance using t tests. Analysis of variance (ANOVA) was used to compare the means of three or more groups. A post hoc test was performed when the ANOVA revealed a significant P value. Data sets with non-normal distributions were log-transformed. A P value < 0.05 was considered significant. To determine the independent association of anemia with various risk factors, adjusted odds ratios (aORs) and 95% CIs were calculated using multivariate logistic regression.

Ethical considerations.

Ethical approval for the study was obtained from the University of California, Irvine, Institutional Review Board (HS no. 2017-3512) and Maseno University Ethics and Research Committee in Kenya (MSU/DRPI/MUERC/00456/17).

RESULTS

There were no significant differences in age or gender distributions of the study populations in the irrigated and non-irrigated areas during the 2018 wet season or 2019 dry season ( Supplemental Table S1). The median age of participants in both the irrigated and non-irrigated areas was 21 years (interquartile range, 44–47 years).

Prevalence of anemia and Plasmodium infection.

Anemia prevalence was 44.8% (319 of 712) during the wet season (June 2018) and 29.5% (244 of 826) during the dry season (February 2019) (χ² = 17.72, degrees of freedom [df] = 1, P value < 0.0001). Table 2 presents data showing that the overall prevalence of anemia and Plasmodium infection during the wet season differed in residents of households located in the irrigated and non-irrigated areas. Anemia prevalence in the irrigated areas was significantly less in the irrigated areas than in the non-irrigated areas (38.9% versus 51.5%, P value < 0.05). A similar trend was observed when age was stratified to 5 to 11 years and ≥ 12 years (but not for children < 5 years). In contrast, Plasmodium infection prevalence was significantly greater in the irrigated are than in the non-irrigated areas (15.3% versus 7.8%, P value < 0.05). The geometric mean density of Plasmodium infection was similar in the irrigated versus the non-irrigated study areas (381 versus 309 parasites/µL, respectively). Infection prevalence for the 5- to 11-year and ≥ 12-year age groups, but not for children < 5 years, was greater in the irrigated areas than in the non-irrigated areas. Given the relative lack of naturally acquired immunity to blood-stage infection in young children in this area of western Kenya,²⁰ we speculate that our study was underpowered to make meaningful conclusions with respect to the impact of residency in irrigated versus non-irrigated areas. As in the wet season, anemia prevalence in the dry season was significantly less in study participants residing in the irrigated areas than in the non-irrigated areas (25.2% versus 34.1%, P value < 0.05). This difference was evident in the two older age groups; but, as in the survey conducted for the wet season, not in children younger than 5 years.

Table 2.

Prevalence of anemia and Plasmodium infection in irrigation and no-irrigation communities

Variable	Irrigation, n/N (%)	No irrigation, n/N (%)
Wet season (June 2018)
Overall anemia prevalence	148/380 (38.9)*	171/332 (51.5)
Age, years
< 5	28/53 (52.8)	34/53 (64.2)
5–11	31/112 (27.7)*	42/88 (47.7)
≥ 12	88/215 (40.9)*	98/191 (50.3)
Overall Plasmodium infection prevalence	58/380 (15.3)*	26/332 (7.8)
Age, years
< 5	5/53 (9.4)	5/53 (9.4)
5–11	25/112 (22.3)*	9/88 (10.2)
≥ 12	28/215 (13)*	12/191 (6.3)
Dry season (February 2019)
Overall anemia prevalence	107/424 (25.2)*	137/402 (34.1)
Age, years
< 5	40/92 (43.5)	39/85 (45.9)
5–11	3/69 (4.3)*	12/55 (21.8)
≥ 12	64/263 (24.3)*	86/262 (32.8)
Overall Plasmodium infection prevalence	7/424 (1.7)	4/402 (1)
Age, years
< 5	1/92 (1.1)	1/85 (1.2)
5–11	1/69 (1.45)	1/55 (1.82)
≥ 12	5/263 (1.9)	2/262 (0.8)

^*P value < 0.05 vs. no-irrigation area.

We also compared the prevalence of anemia according to gender and severity in the wet and dry seasons. With regards to the former, females ≥ 12 years were more commonly anemic than males in both the wet and dry seasons. Females < 12 years, for example, were 38.5% anemic, whereas the males of the same age group were 50% anemic during the wet season (P value = 0.06). In contrast, females ≥ 12 years were significantly anemic (56.9%) compared with males (24.6%, P value < 0.0001). During the dry season, the proportion of anemic females and males < 12 years was 29.4% and 32.7%, respectively (P value = 0.6). However, females ≥ 12 years were more anemic (31.7%) compared with males of the same age group (20.9%, P value = 0.01).

In terms of anemia severity, severe, moderate, and mild anemia accounted for 3.5% (25 of 712), 18.8% (134 of 712), and 22.5% (160 of 712) of the total for irrigated and non-irrigated areas combined during the wet season (χ² = 113.4, df = 2, P value < 0.0001). Similarly, during the dry season, the prevalence of severe, moderate, and mild anemia was 3.1% (26 of 826), 12.3% (102 of 826), and 14.0% (116 of 826), respectively (χ² = 63.97, df = 2, P value < 0.0001).

Mean Hb concentrations.

Children younger than 5 years five had significantly lower Hb concentrations than other age groups in both the irrigated and non-irrigated areas during the wet season. For example, the mean Hb concentration for participants < 5, 5 to 11, and ≥ 12 years was 10.4, 11.8, and 12.4 g/dL, respectively (ANOVA: F = 47.69, df = 2, n = 709, P value < 0.0001). Similarly, during the dry season, the mean Hb concentration among children < 5, 5 to 11, and ≥ 12 years was 10.9, 12.9, and 12.9 g/dL, respectively (ANOVA: F = 67.9, df = 2, n = 823, P value = 0.002).

During wet season, the mean Hb concentration among study participants living in non-irrigated areas was less (mean ± SD, 11.6 ± 2.098 g/dL) than in irrigated areas (12.2 ± 1.832 g/dL, P value = 0.0002). The same trend was observed during the dry season. The mean Hb concentration of non-irrigated-area residents (12.3 ± 2.18 g/dL) was significantly less than that of irrigated-area participants (12.7 ± 2.004 g/dL, P value = 0.006)

Females had significantly lower Hb concentrations (11.7 ± 1.860 g/dL) than males (12.2 ± 2.115 g/dL) during the wet season (P value = 0.002). Similarly, during the dry season, Hb concentration in females (12.3 ± 2.084 g/dL) was significantly less than in males (12.8 ± 2.279 g/dL, P value = 0.0005).

Socioeconomic and nutritional determinants of anemia risk.

Annual household income differed significantly between irrigated and non-irrigated areas (Table 3). The proportion of those earning less than USD365/year was significantly greater in non-irrigated compared with irrigated areas (χ² = 6.31, df = 1, P value = 0.01). The proportion of households earning more than USD1,095/year was significantly greater in irrigated than in non-irrigated areas (χ² = 19.7, df = 1, P value < 0.0001).

Table 3.

Annual household income and food intake per week

Variable	Irrigated (N = 91), n (%)	Non-irrigated (N = 120), n (%)	χ2, df	P value
Household income per year, USD
< 365	8 (8.8)	26 (21.7)	–	–
365–1,095	47 (51.6)	78 (65)	21.3, 2	< 0.0001
> 1,095	36 (39.6)	16 (13.3)	–	–
Iron-rich food (plant)/week
Daily	68 (74.7)	65 (54.2)	–	–
4–6 times	9 (9.9)	16 (13.3)	150.4, 2	< 0.0001
1–3 times	14 (15.4)	39 (32.5)	–	–
Iron-rich food (animal products)/week
Daily	10 (11)	4 (3.3)	–	–
4–6 times	14 (15.4)	11 (9.2)	260.6, 3	< 0.0001
1–3 times	61 (67)	77 (64.2)	–	–
No intake	6 (6.6)	28 (15)	–	–
Vitamin B-rich fruit intake/week
Daily	9 (9.9)	7 (5.8)	–	–
4–6 times	14 (15.4)	14 (11.7)	238.5, 3	< 0.0001
1–3 times	56 (61.5)	80 (66.7)	–	–
No intake	12 (13.2)	19 (15.8)	–	–
Vitamin C-rich fruit intake/week
Daily	7 (7.7)	8 (6.7)	–	–
4–6 times	9 (9.9)	11 (9.2)	132.7, 3	< 0.0001
1–3 times	36 (39.6)	59 (49.2)	–	–
No intake	39 (42.9)	42 (35)	–	–

df = degrees of freedom; USD = U.S. dollars.

Table 4 .

Factors associated with anemia among study participants during the dry season (N = 826)

Variable	Anemia		Crude OR (95% CI)	P value	Adjusted OR (95% CI)	P value
	Yes, n (%)	No, n (%)
Region
Non-irrigated	137 (34.1)	265 (65.9)	1.5 (1.13–2.07)	0.005	1.2 (0.87–1.79)	0.23
Irrigated	107 (25.2)	317 (74.8)	1.00	–	1.00	–
Age, years
< 5	73 (42.7)	98 (57.3)	2.0 (1.42–2.87)	< 0.0001	2.1 (1.13–4.08)	0.02
5–11	15 (12.1)	109 (87.9)	0.34 (0.19–0.61)	< 0.0001	0.35 (0.14–0.89)	0.03
≥ 12	156 (29.4)	375 (70.6)	1.00	–	1.00	–
Gender
Female	158 (31.1)	350 (68.9)	1.22 (0.89–1.66)	0.21	1.4 (0.98–1.92)	0.06
Male	86 (27)	232 (73)	1.00	–	1.00	–
Annual income (USD)
< 365	77 (63.6)	44 (36.4)	22.1 (10.3–47.5)	< 0.0001	9.7 (3.66–26.0)	< 0.0001
365–1,095	164 (36.4)	287 (63.6)	6.0 (3.2–11.1)	< 0.0001	4.4 (2.1–9.1)	< 0.0001
> 1,095	3 (1.2)	251 (98.8)	1.00	–	1.00	–
Iron-rich food intake (plants)/week
1–3 times	62 (35.2)	114 (64.8)	3.7 (2.36–5.70)	< 0.0001	2.8 (1.57–4.84)	< 0.0001
4–6 times	42 (36.8)	72 (63.2)	1.8 (1.02–3.24)	0.4	0.6 (0.27–1.33)	0.22
Daily	140 (26.1)	396 (73.9)	1.00	–	1.00	–
Iron-rich food intake (animals)/week
No intake	77 (58.8)	54 (41.2)	14.8 (5.14–42.5)	< 0.0001	7.1 (1.98–25.4)	0.003
1–3 times	148 (29.7)	350 (70.3)	3.0 (1.14–8.12)	0.03	2.3 (0.76–7.09)	0.1
4–6 times	10 (10)	90 (90)	1.7 (0.54–5.27)	0.4	2.5 (0.7–9.47)	0.2
Daily	9 (9.3)	88 (90.7)	1.00	–	1.00	–
Plasmodium infection
Infected	4 (36.4)	7 (63.4)	1.4 (0.4–4.72)	0.6	1.13 (0.28–4.45)	0.9
Not infected	240 (29.4)	575 (70.6)	1.00	–	1.00	–
Vitamin C intake/week
No intake	213 (67.8)	101 (32.2)	10.2 (4.0–25.8)	< 0.0001	8.7 (2.41–31.27)	0.001
1–3 times	86 (23.1)	287 (76.9)	1.4 (0.57–3.7)	0.4	0.98 (0.28–3.39)	0.97
4–6 times	17 (21.5)	62 (78.5)	1.3 (0.4–4.1)	0.6	0.89 (0.18–4.47)	0.9
Daily	10 (16.7)	50 (83.3)	1.00	–	1.00	–
Vitamin B intake/week
No intake	84 (68.9)	38 (31.1)	8.1 (3.18–20.47)	< 0.0001	4.2 (1.43–12.1)	0.009
1–3 times	192 (36.2)	338 (63.8)	2.06 (0.91–4.66)	0.08	1.6 (0.65–3.90)	0.3
4–6 times	36 (32.7)	74 (67.3)	1.77 (0.69–4.54)	0.2	1.9 (0.68–5.43)	0.2
Daily	14 (21.9)	50 (78.1)	1.00	–	1.00	–

1 = reference value; OR = odds ratio; USD = U.S. dollars.

Nutrition and food intake.

The majority of participants (74.7%) from irrigated areas consumed non-heme iron foods on a daily basis compared with 54.2% from non-irrigated areas (Table 3). Similarly, a greater proportion of residents in irrigated areas consumed heme iron foods 7 days/week versus residents of non-irrigated areas (χ² = 4.9, df = 1, P value = 0.03). The results further indicate that a significant proportion of study participants from non-irrigated areas did not consume heme iron foods when compared with those from irrigated areas (χ² = 10.73, df = 1, P value = 0.001). Although there was a significant difference in vitamin consumption between irrigated and non-irrigated areas, it is clear that few people in both areas consumed vitamins B and C (Table 3).

Factors associated with anemia.

Anemia risk factors were determined during the dry season (second survey). Prior to the start of the study, it was hypothesized that Plasmodium infection would be associated with anemia. During the wet season, however, there was no link between anemia and Plasmodium infection. This prompted us to consider other possible risk factors that were evaluated during the dry season, as indicated in Table 3.

Children younger than 5 years were 2.1 times (aOR, 2.1; 95% CI, 1.13–4.08) more likely to be anemic than those ≥ 5 years. Study participants with a family annual income of less than USD365 and USD365 to 1,095 were 9.7 times (aOR, 9.7; 95% CI, 3.66–26.0) and 4.4 times (aOR, 4.4; 95% CI, 2.1–9.1) more likely to be anemic, respectively, than participants with a family annual income of more than USD1,095.

Study participants who consumed non-heme iron foods one to three times per week were 2.8 times (aOR, 2.8; 95% CI, 1.57–4.84) more likely to be anemic than those who ate non-heme iron foods four to six times/week or on a daily basis. Participants who did not eat heme iron foods during the week were 7.1 times (aOR, 7.1; 95% CI, 1.98–25.4) more likely to be anemic than those who consumed it once a week, one to three times per week, four to six times per week, and on a daily basis. Study participants who did not consume vitamin C were 8.7 times (aOR, 8.7; 95% CI, 2.41–31.27) more likely to be anemic than those who did one to three and four to six times per week and daily.

Participants who did not consume vitamin B were 4.1 times (aOR, 4.2; 95% CI, 1.43–12.1) were more likely to be anemic than those who consumed it for one to three or four to six times daily. There was no association between Plasmodium infection and anemia (aOR, 1.4; 95% CI, 0.28–4.45). Similarly, neither gender nor place of residence was associated with anemia (aOR, 1.4; 95% CI, 0.98–1.92) and (aOR, 1.2; 95% CI, 0.87–1.79), respectively (Table 3).

DISCUSSION

This study was undertaken to compare the rates of anemia, Plasmodium species infection, nutritional intake, and household income in the study area. Anemia is a major health concern in malaria-endemic regions, causing significant morbidity and mortality.²¹ Surveys assessing Hb levels in populations are critical to developing appropriate interventions that may alleviate the severity of anemia. During the wet season, the prevalence of anemia was 44.8%. This rate of anemia is regarded as a severe public health problem per WHO standards.¹⁷ These results are consistent with other study findings from malaria-endemic zones in Kenya where anemia has also been reported to pose a considerable challenge to public health.^13,14,22,23 All of these studies pointed to malaria as the primary cause of Hb deficiency.^13,14,22 However, in our study, there was no association between Plasmodium infections and anemia during the two cross-sectional surveys. The lack of association could be attributed to a significant reduction in parasite density and prevalence. During the study period, an indoor residual spray program using Actellic 300 CS insecticide was implemented in February 2018 and repeated in February 2019. The indoor residual spray program significantly reduced both the Plasmodium infections and parasite density (C. J. Omondi, unpublished data). These findings are consistent with previous research that found no significant association between Plasmodium infection and anemia resulting from low malaria prevalence and parasitemias.^6,24

The Plasmodium infection rate in irrigated areas was greater than in non-irrigated areas. Previous studies in Tanzania,²⁵ Ethiopia,²⁶ and Mali²⁷ reported increased malaria within irrigated compared with non-irrigated areas.

Our study found that males were more likely than females to be infected with Plasmodium. Although some studies reported similar findings,²⁸ others indicated that females were more infected by malaria parasites than males.²⁹ In our study area, some males were engaged in night fishing activities, whereas others were engaged in rice farming, which typically begins very early and ends at dusk. These activities may have exposed males to infectious mosquito bites, resulting in an increase in Plasmodium infections. Children younger than 5 years had the lowest Plasmodium infection rate. This finding is in agreement with previous studies that reported an increase in malaria infections with increasing age.³⁰ It was also important to note that, although the Plasmodium infection rate among children < 5 years was similar in irrigated and non-irrigated areas, the malaria infection rate among children > 5 years was significantly greater in irrigated areas. Older children have been reported to use bed nets infrequently.³¹ Inadequate use of bed nets will almost certainly lead to an increase in infections, particularly in irrigated areas. Furthermore, this infection pattern was only observed during the wet season, when rice farming is at its peak, implying that older children may be involved in rice farming. Despite an increase in Plasmodium infection rates in irrigated areas, anemia cases were significantly less in irrigated areas than in non-irrigated areas. This finding is similar to a study conducted in Tanzania,²⁸ which reported lower anemia rates in areas with greater Plasmodium infection rates.

The variation in anemia rates between the two areas may have been influenced by household income and diet, which was better in irrigated areas than in non-irrigated areas. Low-income households were 4.4 to 9.7 times more likely to be anemic. This is consistent with other studies that reported increased anemia rates in low-income populations.³² A higher income influences the affordability and the frequency with which food can be consumed. Iron intake, both heme and non-heme, was also associated with the prevalence of anemia. Those who could not afford to eat foods rich in heme iron, for example, were 7.1 times more likely to be anemic. Similarly, those who could only consume foods rich in non-heme one to three times per week were 2.8 times more likely to be anemic than those who consumed it on a daily basis. A study conducted in northeastern Ethiopia also reported that children with poor dietary intake were more likely to be anemic compared with those who ate a healthy diet.²⁵ Although there are many underlying factors that may lead to anemia, a poor diet and low household income were some of the possible causes of anemia in the study area.

Our study reported a generally poor intake of vitamins B and C in both irrigated and non-irrigated areas. Vitamin C is an essential nutrient that aids in the absorption of non-heme iron.³³ In both surveys, moderate and mild anemia, which most often occurs silently within populations, were significantly greater compared with severe cases. This finding is in agreement with other studies.³⁴ It is also worth noting that the prevalence of anemia declined significantly during the dry-season survey. Reduced malaria cases during the dry season may have contributed to the observed reduction in anemia rates.

In both surveys, the prevalence of anemia prevalence varied significantly by age group. The majority of anemia cases were observed in children younger than 5 years. It is noteworthy that this age group had the lowest prevalence of malaria infections. Previous studies also reported the same.^32,34,35 Young children, in general, experience rapid physical and mental development, which requires an adequate supply of micronutrients such as iron, vitamin B₁₂, and folic acid for optimal growth.³⁶ Anemia in actively growing children may be caused by the lack of or insufficient supply of these nutrients. Anemia has a negative and long-term impact on children’s mental growth,³⁷ it retards normal body development,³⁸ and it affects their behavior.³⁹ Furthermore, our results show a progressive decline in anemia severity with age, with adults having greater rates of mild anemia than younger age groups. This is consistent with other research findings, which show that as one gets older, the severity of anemia decreases.^13,24,40,41 It has been suggested that as the immune system matures with an increase in age, increased synthesis of hemoglobin is enhanced.^42,43 This can explain in part the declining trend of anemia toward adulthood. Previous research suggests that frequent exposure to Plasmodium infections in adults may boost the immune system against malaria, resulting in a reduction in anemia.⁴⁴

Males had greater anemia rates than females among participants < 12 years. This is consistent with previous research.⁴⁵ According to one study,⁴⁵ males have greater levels of testosterone before puberty, which stimulates rapid growth and thus an increased demand for iron than females. However, from 12 years onward, anemia was found to be more prevalent in females than in males. This is also in agreement with other previous findings.⁴⁶ The onset of menstruation, coupled with rapid growth in adolescent females, are considered major causes of anemia in this age group.^2,47 There is a greater demand for iron, which is lost during menarche, and there is a faster growth rate. According to previous studies, females from low-income families are more likely to develop anemia as a result of an unhealthy diet.⁴⁸

One of our study’s limitations was its inability to investigate the morphological appearance of red blood cells. This made it impossible to distinguish between anemia caused by Plasmodium and other forms of anemia, such as iron deficiency anemia, sickle cell anemia, and anemia caused by other parasitic infections. Because of the cross-sectional nature of the study, determining the causal relationships between anemia and independent variables was also difficult.

CONCLUSION

Anemia is a severe public health problem, particularly in non-irrigated areas. Anemia was most prevalent in children younger than 5 years and in females 12 years and older. Anemia was associated significantly with age, household income, iron-rich food consumption, and vitamin consumption. As a result, poverty alleviation and proper dietary intake are critical to reducing the anemia burden in this region. To guide policymakers on how to mount expanded and targeted control intervention measures, detailed studies revealing causal links are required. It is noteworthy that the type of irrigation practiced in our study contributed to better nutrition that mitigated against anemia in irrigated areas. Intensive vector control may have reduced the impact of malaria infections on anemia in both the irrigated and non-irrigated areas.

Note: Supplemental table appears at www.ajtmh.org.