Background
Diarrhoea is the second leading cause of death in children under five years [
1]. Diarrhoea can be caused by bacteria, viruses, and parasites [
2‐
4]. Among parasites, the intestinal protozoa
Cryptosporidium spp. is the most attributable to moderate-to-severe diarrhoea in children younger than five in sub-Saharan Africa and Asia [
3]. The second and third most common intestinal protozoa identified in children with diarrhoea are
Giardia lamblia and
Entamoeba histolytica, respectively [
5‐
7]
.
Transmission of intestinal protozoa occurs by the ingestion of contaminated food and water or by direct contact with infected individuals or animals [
8,
9]. Host and environmental characteristics such as age, nutritional status, access to treated water, animal contact, and, population density are important factors in the occurrence of parasites infections [
10‐
14].
Mozambique is a low-income country, with reports of intestinal parasites infections in children with diarrhoea, mostly in the southern region, having identified a frequency of 16.1% in Maputo city and 14.4% in Manhiça district, using microscopic diagnostic techniques [
3,
7,
15,
16]. Using a more sensitive technique, the Enzyme-Linked Immunosorbent Assay (ELISA), a second study also in the Manhiça district reported the frequency of
Cryptosporidium spp.
, G. lamblia, and
E. histolytica/dispar in children with diarrhoea by 18.7, 17.2, and 10.2%, respectively [
7].
A national survey in school-age children identified Nampula - a northern province of Mozambique - as one of the most affected by parasites infections [
17]. A second study conducted in
Hospital Central de Nampula (HCN) observed
G. lamblia as the most common parasite, followed by
Cryptosporidium spp. in a pooled estimation in children younger than five years with either diarrhoea, malnutrition, or Human Immunodeficiency Virus (HIV) [
18]. However, parasites infections in children with only diarrhoea are not reported, neither risk factors for this group [
18].
A pooled analysis of a surveillance system in four provinces of Mozambique indicated Nampula province as the one with the highest risk of infection for
Cryptosporidium spp. in children younger than 15 years [
5]. Nampula province has poor health indicators, with recurrent reports of cholera outbreaks [
19]. The previous results from the surveillance system (
Vigilância Nacional de Diarreias – ViNaDia or National Surveillance of Diarrhoea) failed to indicate risk factors for protozoans in each province [
5]. Therefore, this analysis aims to determine the frequency of intestinal protozoa infections and associated risk factors in children younger than five years admitted with diarrhoea in
Hospital Central de Nampula, in the Northern region of Mozambique.
Results
During the analysis period, 276 children admitted at HCN were included. However, 99.3% (274/276) of the children provided sufficient stool amounts to obtain results for Cryptosporidium spp. and G. lamblia and, 99.6% (275/276) for E. histolytica. Overall, 98.9% (273/276) of the children had enough stool amounts to obtain results for the three protozoans.
Sample characteristics
Among the included children, 55.8% (154/276) were male; 48.6% (134/276) were less than 12 months of age; 46.0% (127/276) had caregivers with a secondary or greater educational level. Also, 64.5% (178/276) were living in a household with more than four members (Table
1). Animal contact was reported in 46% (127/276); 51.1% (141/276) drank water from the public tap, and 63.6% (171/276) were well-nourished for WAZ (Table
1).
Table 1
Characteristics of children hospitalized with diarrhoea in HCN, March 2015 to January 2018, Nampula, Mozambique
Sex |
Male | 154 | 55.8 |
Female | 122 | 44.2 |
Age categorized (in months) |
0–11 | 134 | 48.6 |
12–23 | 93 | 33.7 |
24–59 | 49 | 17.8 |
Caregiver Education Level |
Illiterate | 42 | 15.2 |
Primary | 104 | 37.7 |
Secondary/Above | 127 | 46.0 |
Unknown/Missing | 3 | 1.1 |
Number of Members in the Child Household |
< 5 | 76 | 27.5 |
≥ 5 | 178 | 64.5 |
Unknown/Missing | 22 | 8.0 |
Animal contact |
No | 148 | 53.6 |
Yes | 127 | 46.0 |
Unknown/Missing | 1 | 0.4 |
Drinking-Water Source |
Well | 58 | 21.0 |
Piped water | 74 | 26.8 |
Public tap | 141 | 51.1 |
Others | 2 | 0.7 |
Unknown/Missing | 1 | 0.4 |
Underweight (Weight-for-age Z-Score) |
No | 171 | 63.6 |
Yes | 79 | 29.4 |
Unknown/Missing | 19 | 7.1 |
The overall frequency of intestinal protozoa
Infection by any enteric protozoa was seen in 21.4% (59/276; 95% CI: 17.0–26.6) of the children. Cryptosporidium spp. was the most common protozoa at 13.9% (38/274; 95% CI: 10.3–18.5), followed by G. lamblia at 9.1% (25/274; 95% CI: 6.3–13.1) and E. histolytica at 0.4% (1/275; 95% CI: 0.1–2.0).
Positive children for Cryptosporidium spp. and G. lamblia had a median age (in months) and interquartile interval of 15 (10–24) and 11 (8–20), respectively.
Co-infection between two protozoa was observed in 1.8% (5/273) of children. The most and least common co-infections were Cryptosporidium spp./G. lamblia 1.4% (4/273) and, Cryptosporidium spp./E. histolytica 0.4% (1/273), respectively.
Children whose caregivers had primary education level were less likely to be infected by
Cryptosporidium spp. (Adjusted Odds Ratio: 0.290, 95% CI: 0.091–0.916) compared with the illiterates. Underweighted children were 2.7 times more likely of being infected by
Cryptosporidium spp. compared to well-nourished children (p-value = 0.011; 95% CI: 1.261–5.798) (Table
2).
Table 2
Characteristics, frequencies, crude and adjusted odds ratio for infection by Cryptosporidium spp. and Giardia lamblia, in children hospitalized with diarrhoea in HCN, March 2015 to January 2018, Nampula, Mozambique
Sex |
Male | 153 | 16 (10.5) | | 1 | | 1 | 14 (9.2) | 0.986 | 1.007 (0.440–2.306) | | |
Female | 121 | 22 (18.2) | 0.069 | 1.903 (0.951–3.809) | 0.045 | 2.157 (1.017–4.572) | 11 (9.1) | | 1 | | |
Age categorized (in months) | | | 0.642 | | | | | 0.173 | | 0.332 | |
0–11 | 134 | 21 (15.7) | 0.375 | 1.598 (0.567–4.506) | | | 8 (6.0) | | 1 | | 1 |
12–23 | 92 | 12 (13.0) | 0.652 | 1.290 (0.426–3.903) | | | 10 (10.9) | 0.187 | 1.921 (0.728–5.069) | 0.227 | 1.834 (0.686–4.904) |
24–59 | 48 | 5 (10.4) | | 1 | | | 7 (14.6) | 0.071 | 2.689 (0.919–7.870) | 0.191 | 2.224 (0.671–7.367) |
Caregiver Education Level | | | 0.075 | | 0.042 | | | 0.774 | | | |
Illiterate | 42 | 8 (19.0) | | 1 | | 1 | 4 (9.5) | | 1 | | |
Primary | 103 | 8 (7.8) | 0.056 | 0.358 (0.15–1.028) | 0.035 | 0.290 (0.091–0.916) | 11 (10.7) | 0.834 | 1.136 (0.340–3.791) | | |
Secondary/Above | 126 | 22 (17.5) | 0.816 | 0.899 (0.367–2.205) | 0.803 | 0.883 (0.331–2.351) | 10 (7.9) | 0.748 | 0.819 (0.243–2.763) | | |
Unknown/Missing | 3 | – | | | | | – | | | | |
Number of household member |
< 5 | 76 | 12 (15.8) | | 1 | | | 2 (2.6) | | 1 | | 1 |
≥ 5 | 176 | 21 (11.9) | 0.406 | 0.723 (0.336–1.555) | | | 21 (11.9) | 0.032 | 5.013 (1.145–21.947) | 0.039 | 4.762 (1.083–20.935) |
Unknown/Missing | 22 | – | | | | | – | | | | |
Animal contact |
No | 146 | 24 (16.4) | 0.200 | 1.588 (0.783–3.220) | 0.091 | 1.960 (0.899–4.273) | 12 (8.2) | | 1 | | |
Yes | 127 | 14 (11.0) | | 1 | | 1 | 13 (10.2) | 0.565 | 1.273 (0.559–2.901) | | |
Unknown/Missing | 1 | – | | | | | – | | | | |
Drinking-Water Source | | | 0.373 | | | | | 0.855 | | | |
Well | 56 | 5 (8.9) | | 1 | | | 5 (8.9) | 0.925 | 1.054 (0.353–3.142) | | |
Piped water | 74 | 10 (13.5) | 0.421 | 1.594 (0.512–4.957) | | | 8 (10.8) | 0.582 | 1.303 (0.508–3.344) | | |
Public tap | 141 | 22 (15.6) | 0.225 | 1.886 (0.677–5.255) | | | 12 (8.5) | | 1 | | |
Others | 2 | 1 (50.0) | 0.119 | 10.200 (0.550–189.125) | | | 0 (0.0) | | – | | |
Unknown/Missing | 1 | – | | | | | – | | | | |
Underweight (Weight-for-Age Z-Score) |
No | 170 | 19 (11.2) | | 1 | | 1 | 16 (9.4) | | 1 | | |
Yes | 78 | 17 (21.8) | 0.030 | 2.215 (1.079–4.544) | 0.011 | 2.704 (1.261–5.798) | 9 (11.5) | 0.606 | 1.255 (0.529–2.980) | | |
Unknown/Missing | 19 | – | | | | | – | | | | |
Children living in a household with more than four members were 4.7 times more likely to be infected by
G. lamblia compared with children from households with fewer members (
p-value = 0.039; 95% CI: 1.083–20.935) (Table
2).
E. histolytica was detected in a single male child, 11 months old, co-infected by Cryptosporidium spp. and undernourished. Due to the low positivity, logistic regression models were not built for infection by E. histolytica.
Discussion
Findings from this analysis suggest that pathogenic intestinal protozoa infection is a public health problem in this setting, with one in five children with diarrhoea infected with at least one intestinal protozoa. The overall frequency of parasitic infections (21.4%) is higher than reported in the southern region of Mozambique (14.4 and 16.1%) [
15,
16]. Differences observed may be due to the fact we used a more sensitive technique compared to microscopy used in the previous studies [
27,
28].
Cryptosporidium spp. was the most common parasite (13.9%), followed by
G. lamblia (9.7%) and
E. histolytica (0.4%). Using the same diagnostic technique, a rural setting in the Manhiça district showed higher frequencies for
Cryptosporidium spp. (18.7%) and
G. lamblia (17.2%) among children with diarrhoea [
7].
E. histolytica was observed in less than 1 % of the overall samples, similar to the reported in Tanzania, a country near Nampula province [
6]. However, in the Manhiça district, the frequency of
E. histolytica/dispar was much higher, 10.2%, suggesting different geographic distribution for this protozoa [
7]. Specific tools for
E. histolytica would better estimate of the occurrence of this parasite in the Manhiça district as the diagnostic tool used detects
E. histolytica/dispar. The Manhiça district is a rural setting, where the children have more contact with soil, unimproved sanitary conditions including management of sewage, hygienic habits, and access to parasitic drugs [
29].
E. histolytica was detected in an eleven months male child, underweighted and co-infected by
Cryptosporidium spp. It is suggested that the male gender is more susceptible to infections due to androgens which reduce host immunity [
30].
Less than 2% of the sample were co-infected. Previous studies conducted elsewhere in Africa also observed co-infections ranging from 0.25 to 43.4% [
31‐
36]. It has been postulated that co-infections with
G. lamblia can modulate the presence of symptoms in enteric infections. For example, co-infection between
G. lamblia and rotavirus showed a fewer number of diarrhoea episodes compared to single infections by rotavirus [
37]. On the other hand, previous studies indicated that
G. lamblia was more common in non-diarrhoeal children [
3,
7,
38], suggesting that symptoms in children infected by
G. lamblia will be observed if there is another enteric pathogen that can contribute to a clinical profile. Further analysis needs to be done to evaluate the clinical characteristic of co-infections, including the presence of
G. lamblia.
The caregiver’s education level was an independent risk factor for infection by
Cryptosporidium spp. It was previously observed that a higher education level is directly associated with better hygiene habits and sanitary conditions. Therefore children with caregivers who are literate will be less susceptible to infection [
10‐
12,
39,
40].
Undernutrition and infection by
Cryptosporidium spp. share a similar geographic distribution [
13]. The infection can cause inflammation and damage to the small intestine, leading to an undernourished profile and growth impairment [
13]. An association between
G. lamblia and nutritional status was not observed in this analysis. Although this relationship has been reported for chronic infection with
G. lamblia [
41]. However, we could not identify if children positive for
G. lamblia had a chronic infection.
A high number of members in the children’s household was a risk factor for infection by
G. lamblia. Little is known regarding this outcome, especially when protozoa infective-dose is considered; however, it is known that crowded environments contribute to parasites transmission [
11,
12]. On the contrary, the relation between
Cryptosporidium spp. and the number of household members was not observed.
G. lamblia was more common in children without diarrhoea aged 12 to 59 months than in the ones with diarrhoea in the Global Enteric Multicenter Study (GEMS), which included Gambia, Mali, Mozambique, Kenya, India, Bangladesh, and Pakistan. Therefore, it seems that
G. lamblia is more ubiquitous than
Cryptosporidium spp., making the probability of a child to be infected by
G. lamblia in a crowded environment much greater than by
Cryptosporidium spp. [
3].
A previous study conducted at HCN between 2012 and 2013, included children with undernourished, HIV, and diarrhoea, observed
G. lamblia genotype B as the most common, which is related to anthroponomical transmission [
18]. Previous results from HCN suggests that the absence of relation between animal contact and infection by
G. lamblia in our sample may indicate the circulation of an anthroponomical genotype [
18].
The drinking water source has been suggested as one of the routes for transmission of intestinal protozoa [
14,
42,
43]. Transmission by water sources has been pointed mostly in developed countries outside Africa. Detection and notification are mostly reported by water suppliers in the outbreak context [
9]. No outbreak was recorded during the survey period in the HCN, which can justify the absence of a relationship between drinking water source and infection by an intestinal protozoa.
The limitations in this analysis were: collection of a single stool sample from each participant, although it is recommended testing three samples for each individual, to avoid underestimation; and using a cross-sectional study design, which cannot measure causality between dependent and independent variables [
44,
45]. However, we used a more sensitive diagnostic technique to minimize the underestimation of the true occurrence of intestinal protozoa in our samples [
27,
28].
Our data indicate the presence of intestinal protozoa in 21.4% of the children with diarrhoea in HCN, however, 78.6% of the hospitalized children have unknown aetiology. Novel diagnostic tools such as multiplex techniques that identify a wider range of enteric pathogens should be considered to estimate the true burden of enteric pathogens in children with diarrhoea as well as co-infections and their clinical profiles especially in the presence of
G. lamblia [
4,
37].
Acknowledgments
We want to thank the caretakers who consented for their children to be enrolled in the surveillance. For their contribution in the ViNaDia procedures, we would like to thank Herminio Cossa, Juma Cantoria, Osvaldo Laurindo, Angelina Pereira, Mulaja Kabeya Étienne, Celso Gabriel, Titos Maulate, Julieta Ernesto, Saide Artur, Musangu Mukinayi, Francisca Ricardo, Siasa Mendes, Marcelino Tsowo, Direcção Clínica de Pediatria do Hospital Central de Nampula, Carlos Guilamba, Celina Nhamuave and Miguel Bambo.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.