Prevalence and risk factors of human Balantidium coli infection and its association with haematological and biochemical parameters in Ga West Municipality, Ghana

Background In Ghana, Balantidium coli (B. coli) has been identified in vegetables and in pigs, although there is a paucity of data regarding human balantidiosis. This study sought to assess human B. coli infection in Ghana, factors associated with the infection as well as its association with haematological and biochemical parameters. Methods Two pig rearing communities in the Ga West Municipality, Ghana, were involved in this study. Stool and blood samples were collected from pig farmers and their exposed household members as well as relevant information on potential associated factors. Eosin-saline wet preparation was done on the same day of stool samples were collected while formol ether concentration technique was performed later. Haematological, biochemical parameters and serum electrolytes were determined using Celltac MEK-6500 K, PKL-125 biochemical analyser, and FT-320 electrolyte analyser, respectively. Results The overall prevalence of balantidiosis was 10.4 %, significantly higher among farmers (21.7 %) than in exposed household members (5.8 %) (x2 = 17.8, p = 0.000025). Of the 43 infected individuals, 20.9 % were co-infected with either Entamoeba histolytica, Giardia lamblia, or Schistosoma mansoni. In B. coli infection, mild to moderate anaemia together with a reduction in levels of platelet, albumin and, sodium, chloride, and bicarbonate ions were observed. However, white blood cells were significantly elevated in infected states. Poor farming practices such as free-range systems, improper disposal of pig faeces, lack of use of protective farming clothing, and unavailability of dedicated farming clothing were found to be associated with B. coli infection status. Finally, frequent diarrhea (OR = 12.30, p = 0.006) with occult blood (OR = 25.94, p < 0.0001) were found to be predictors of B. coli infection. Conclusions Human balantidiosis is endemic in Ga West Municipality, Ghana. Individuals living closed to pig rearing communities presenting with frequent diarrhea with occult blood in stool should be screened and treated for balantidiosis to mitigate the clinical consequences of the infection.

in commonly consumed vegetables in Ghana. Water supply containing pig faeces used for irrigation being the main source of contamination [1]. In the Upper East region of Ghana, almost 20 % of pigs were found to harbour B. coli [2]. The presence of parasites in vegetables and in pigs could serve as a source of human infections.
B. coli is known to be transmitted to humans by domesticated pigs [3], where transmission occurs by ingestion of cysts through direct or indirect contact and through other means such as contaminated food and water by human and pig faeces [4]. After encystment in the intestines, the motile trophozoites attack the intestinal epithelium, creating ulcers. B. coli is able to cause ulcers by secreting an enzyme called hyaluronidase. This enzyme is capable of degrading intestinal tissues and facilitates penetration of the mucosa. B. coli causes bloody diarrhea which is similar to that of amoebic dysentery [5]. Some cardinal symptoms of balantidiosis in humans include passing of loose stools, anorexia, fever, and mild abdominal pain [6]. Currently, Centre for Disease Control has recommended the use of tetracycline, metronidazole, and iodoquinol as the treatment option for human balantidiosis [7].
In Ghana, pigs are reared in many parts of the country for nutritional and economic purposes. Due to the lack of advanced technology in rearing pigs in Ghana and other parts of Africa, most farmers employ the services of household members and other community members to support pig production. Most piggeries are done in sheds and stalls, whereas others are free animals with some co-habiting with humans. Feeding is done manually and droppings were used as manure or indiscriminately discarded [8]. These practices are consistent with factors associated with increased transmission of B. coli [3]. However, there is limited data regarding balantidiosis in Ghanaians. Even though the global prevalence of balantidiosis is very low (0.02-1 %) [9], the clinical correlates of this neglected tropical/subtropical infection must be established. In furtherance to the foregoing, the objective of this study was to assess human B. coli infection and associated clinical factors in Ghana, a sub-Saharan country located in West Africa.

Study design, study sites, and period of sampling
This purposive cross-sectional study was conducted between February-May, 2020. This study was carried out in two pig rearing communities, Opah and Onyansana, two villages which are farming communities in the Ga West Municipality in the Greater Accra region of Ghana. A recent parasitological survey reported 1.6 % prevalence of balantidiosis in these two communities [10].

Selection of study participants
Individuals engaged in pig farming as well as their household members were recruited in this study. In each study area, with the help of community-based surveillance volunteers (CBSV), all pig farmers were identified. The pig farmers then led the research team to their houses to interact with their exposed household members. An exposed household member was defined as anyone who did not actively take part in pig farming but frequently entered the pigsty for one reason or the other as well as those that came into contact with pig droppings and pig faeces soiled farming accoutrement.

Sample size estimation
Using the Cochrane's formula, n = z 2 p(1-p)/d 2 , where n is the sample size, z is the confidence level at 95 % (standard value of 1.96), d is the margin of error at 5 % (standard value of 0.05), the sample size was 384, assuming the prevalence of balantidiosis in pig farming communities was 50 % [11].

Inclusion and exclusion criteria
Individuals included in this study were pig farmers and exposed household members. Individuals negative for B. coli were randomly selected as a noninfected comparative group in a ratio of two noninfected individuals to one infected individual. Household members that did not support farming activities, declined blood sample collection, or did not meet the definition above were not included.

Stool and blood sample collection and relevant information collection
For both pig farmers and their household members, a minimum of 2 g of early morning stool sample was collected as well as whole blood samples in both K 2 -EDTA (~ 3 mL) and clot activator tubes (~ 4 mL). Stool samples were kept at ambient temperature while whole blood samples were kept in coolers containing ice packs prior to arrival in the Ga North Municipal Hospital Laboratory located in Ofankor-Accra. Both structured questionnaires and personal observations were used to elicit responses on the type of housing for the pigs, disposal system for pig droppings, disinfection protocol after leaving the pigsty, use of protective equipment (boots, gloves, face shields and goggles), availability of dedicated farming clothing, and history of deworming the pigs. Enrolees were also asked whether there were restricted entry protocols in place, whether farmers and their household members got soiled with pig droppings when cleaning the pigsty. Information regarding the number of times the farmers enter the pigsty and the frequency of washing farming clothing were also collected.

Physical examination and detection of occult blood in stool samples
Stool samples were physically examined for consistency based on the amount of water present in the stool. Formed specimens were stool samples with solid appearance (with little amount of water), semi-formed specimens were solid samples with adequate amount of water, while the loose specimens did not have visible water but able to spread to take the shape of its container. Mucoid specimens contained visible mucus while blood was observed in bloody specimens. Occult blood was determined by using the blood determinant portion of Uritest 10E Urine reagent strip (URIT Medical Electronics Co. Ltd, Guangxi, China) following the manufacturer's procedure.

Processing of stool samples and analysis of blood specimens
On the same day stool samples were collected, direct wet preparation using eosin saline was performed. The rest of the stool sample was fixed in 10 % buffered formalin. Additionally, haematological parameters were determined using Celltac MEK-6500 K automated analyser (Nihon Kohden, Japan). Serum and plasma samples were stored at − 30 °C until analysis. Serum electrolytes (sodium ions, potassium ions, and chloride ions) were measured by FT-320 electrolyte analyzer (Fortune Co. Ltd, China), while plasma urea, creatinine, ALP, total protein, albumin, AST, ALT, indirect bilirubin, and direct bilirubin were measured by PKL-125 (Paramedical Co. Ltd, Italia) fully automated chemistry analyser. For all the biochemical analysis, Medsource reagents (Ozone Biomedicals Pvt Ltd, Haryana, India) were used. Serum bicarbonate ions were measured spectrophotometrically based on the principle that bicarbonate ions react with phosphoenolpyruvate catalysed by phosphoenolpyruvate carboxylase to form oxaloacetate and phosphate. The oxaloacetate was then converted to malate by the action of malate dehydrogenase and reduced nicotinamide adenine dinucleotide (NADH). The decrease in absorbance at 415 nm resulting from the oxidation of NADH was proportional to the amount of bicarbonate in the sample.

Morphologic characterization of B. coli Detection of B. coli using eosin-saline mixture
Stool samples were emulsified into a uniform mixture with normal saline using a wooden spatula. A drop of the preparation was placed on a microscope slide and emulsified with saline and eosin solution (10 % v/v). The emulsified stool sample was examined using a bright field microscope (objectives ×10 and ×40). Viable parasites were identified by their characteristic shape and motility, while B. coli cyst was identified by its spherical or slightly ovoid shape.

Detection of B. coli cysts using the formol ether concentration (FEC) method
Stool samples were vortexed into a homogenous mixture; then, 1 mL of the homogenate was sieved using nonabsorbent gauze. The sieved homogenate was centrifuged at 3,000 rpm after adding 7.0 mL of formol saline reagent and 3 mL of diethyl ether (Honeywell, USA: bp: 34.6 °C, mp: − 116.3 °C, mm: 74.12 g/mol). Supernatant was discarded and the pellet mixed with iodine solution. The pellet was examined with a bright field microscope, using objectives ×10 and ×40. Parasite cysts were identified by their characteristic spherical or slightly ovoid shape.

Data analysis
Prevalence of B. coli was expressed as the ratio of the number of infected individuals in a given category and the total number of individuals examined in that category. Fisher's exact test was used to compare B. coli infection status between stool parameters (stool consistency and stool occult blood) and history of diarrhea. One-way ANOVA was used to compare haematological and biochemical parameters between B. coli mono-infected, B. coli co-infected, and noninfected individuals. In case of significant difference, Tukey post hoc multiple analysis was performed to determine the specific groups significantly different or similar. Univariate logistic regression analysis was used to determine predictors of B. coli infection. Data were analysed by SPSS Version 24 (Chicago, IL, USA). P-values less than 5 % were considered statistically significant.

Fig. 2 Age range distribution of the study participants
However, B. coli co-existed with only nine of the other parasites (6 E. histolytica and 3 G. lamblia).

B. coli infection status according to biophysical properties of stool sample
Among the various stool sample consistencies, semiformed samples (p = 0.037), loose samples (p < 0.001), and bloody stool samples (p = 0.03) were significantly associated with B. coli infection. Additionally, occult blood in stool was significantly higher among individuals infected with B. coli (x 2 = 68.1, p < 0.05) and whose history of diarrhea was infrequent (x 2 = 9.3, p = 0.002) ( Table 2).   Table 3). Of the 26 infected pig farmers, 88.5 % did not disinfect themselves after leaving the pigsty while the majority of them did not use protective boots and gloves (73.0 and 88.5 %, respectively) while attending to the pigs. None of the infected farmers used a face mask as well as a goggle while attending the animals. Additionally, the farmers did not deworm the pigs (Table 4).

Farmers practices encouraging transmission of balantidiosis
Of the 24 pigsties, entry was restricted to only 12.5 %; most of the farmers infected with B. coli entered the pigsties twice a day (65.3 %) while others entered the pigsties once (19.2 %), thrice (11.5 %), and only one reported entering the pigsty an average of four times a day (3.8 %). Even though the farmers reported getting soiled with pig droppings in the pigsties, 19.2 % and 57.7 % of the  farmers did not wash or wash their farming clothing only once in a week, respectively. It was also observed that 42 % of people that were not farmers came into contact with the farmers clothing as both spouses and children were involved in washing the farmers' clothing. Interestingly, 88.5 % of either the farmers or their close contacts washed the farming apparel with their bare hands. Surprisingly, only one farmer indicated disinfecting clothes with domestic bleach prior to washing and only 11.5 % use protective gloves to wash the soiled clothing ( Table 5). Table 6 reports the various biomarkers analysed in this study. One-way analysis of variance (ANOVA) was used to test the differences in the mean levels of biomarkers analysed in both B. coli mono-infection and co-infection with other parasites, while B. coli negative individuals were used as the noninfected comparative group. In both mono-and co-infections, significant differences were observed in haemoglobin (p = 0.088), red blood cells (p = 0.033), mean cell volume (p = < 0.001), total white blood cells (p = 0.004), lymphocytes (p = 0.002), platelet count (p = 0.010), sodium ions (p = 0.02), chloride ions (p = 0.018), bicarbonate ions (p = 0.025) and albumin (p = 0.004) levels. Haemoglobin levels were found to be significantly reduced in both mono-and co-infections, co-infection levels being significantly lower than monoinfection cases. It was also observed that red blood cell levels were significantly lower among co-infected individuals than in mono-infected ones. However, mean cell volume (MCV) was significantly elevated in co-infection compared to mono-infection. Again, it was observed that total white blood cells were significantly higher in infected cases than the noninfected group. Post hoc analysis indicated significantly higher levels of total white blood cells in mono-infections than co-infections. Lymphocyte levels were significantly elevated in co-infection but not in mono-infections, while platelet levels were significantly reduced in co-infection. Furthermore, sodium and chloride ions were reduced in the infection state, but mono-infection levels did not differ significantly from coinfection levels. Bicarbonate ions were also reduced in the infection state with co-infection levels significantly lower than mono-infection. Albumin levels were also found to be significantly lower in the infection state, while coinfection levels did not differ from mono-infection levels.

Discussion
Balantidiosis is a neglected tropical disease, prevalent in rural communities where pig rearing is common [3]. This study reports the findings from the human balantidiosis survey in Ghana and reports for the first time the burden of human balantidiosis in two pig rearing communities in Ghana as well as the identification of probable at-risk groups. More importantly, the clinical attributes of balantidiosis in humans have been documented. In 2008, a study carried out in the Eastern Region of Ghana reported 13.6 % prevalence of B. coli in commonly consumed vegetables [1]. Cases of human balantidiosis have been reported in some tropical and subtropical countries. Indeed, human balantidiosis prevalence of 28, 29, and 12 % were reported in Papua New Guinea, Bolivia and Venezuela, respectively [9,12,13]. In addition, in China [14], Peru [15], and India [16], cases of balantidiosis in human were described. However, in Ghana, a cross-sectional study, carried out in 2020, in the Ga West Municipality in the Greater Accra Region of Ghana identified asymptomatic B. coli infection in nine individuals in Opah and Onyansana communities [10]. Therefore, this study was purposively designed to unearth the real burden of infection in these two communities. A high prevalence of human balantidiosis (10.4 %) was found in the Opah and Onyansana communities. The prevalence of B. coli infection was similar in the two study sites as well as among the age classes and between males and females. Similar rates of infection were also observed in participants with different educational levels, even though the risk of infection was insignificantly higher in participants with no formal education. However, infection status was significantly associated with pig farmers even though some exposed household members were infected. This observation confirms earlier hypotheses that pig rearing is a high-risk venture for B. coli transmission in humans [17]. Also, males and individuals without formal education were mostly infected, likely because more males than females are normally engaged in various farming activities and are mostly within the active working force [18][19][20]. Moreover, individuals without formal education were less informed about healthy lifestyles and lacked appreciation of control and prevention of parasitic diseases [21]. Although significant number of the farmers and their exposed household members were found to be infected by B. coli and other parasites, the wet preparation and formol ether concentration techniques could exhibit low sensitivity. Saline wet preparation is also inefficient in detecting low levels of parasites [22], while in the case of the concentration technique, the gauze used for filtration could trap the B. coli cysts resulting in false negative in cases of low cyst count. In this study, poor and unhealthy farming practices were reported and could favour transmission of the parasite among farmers and their exposed household members. Indeed, the majority of the farmers neither washed their farm clothing regularly nor decontaminate their clothing prior to washing them. Moreover, almost 70 % of the farmers did not have dedicated farm clothing, thus increasing the risk for several clothes to be contaminated by pig faeces and subsequently contaminate other items in the household. Moreover, some of the pig farms operate either free-range or partially confined systems. This poor husbandry practice could contaminate the environment with infected pig faeces and subsequently facilitate the dissemination of B. coli parasites to individuals that are not directly involved in pig rearing in the community. In this study, these poor husbandry practices led to the infection of almost 6.0 % of the exposed household members with B. coli. Moreover, free-range pigs are more infectious that caged animals [23], thus representing a greater risk to the community when they are allowed to wander. The study also observed that there was no restricted entry into the majority of the pigsties and household members handled soiled farm clothing with bare hands without proper training or coaching on how to hygienically wash contaminated protective clothing. In a previous study, living close to pig farms and contamination of wells, streams, rivers, ponds, and soil with B. coli cysts have been reported as factors associated with the infection [23]. Moreover, B. coli cysts are resistant to sodium hypochlorite, a common bleach used to decontaminate clothing [24].
Regarding the clinical features of balantidiosis, the reduction in haemoglobin and red blood cells as observed in this study is widespread in most pathogenic entero-protozoa infections [25]. Even though bloody stool is common with balantidiosis [26], in this study only about 12 % of the B. coli infection presented with visible blood in stool samples. Additionally, occult blood was detected in all infected stool samples. Visible blood was not observed in almost all infected stool samples, possibly because the infected individuals did not exhibit fulminant infection, which culminates in extensive inflammation of the colon. Reduction in haemoglobin and red blood cell levels together with increased mean cell volume was suggestive of megaloblastic anaemia mostly caused by deficient of vitamin B-12 and / or folate. Similar findings were observed in B. coli infected patients in Venezuela [27] and recently in Thailand [28].
Significant thrombocytopenia was evident in B. coli infected individuals which was worsened in co-infection state. Furthermore, a reduction in levels of sodium, chloride and bicarbonate ions was observed in the infected individuals as was previously reported by Yu et al. [14]. In that case report, mild anaemia, hypokalemia, hyponatremia, and hypochloremia were observed as well as occult blood in stool samples. Reduction in these ions could be as a result of diarrhea experienced by all infected individuals, while the presence of occult blood in all studied individuals could be as a result of parasite-induced colitis [14]. Reduction in albumin levels confirms the association of malnutrition with balantidiosis, as highly documented in several parasitic studies [9,[29][30][31].
This study did not screen the infected individuals and noninfected comparative group individuals for haemoparasites (e.g., malaria, babesia, filarial worms etc.), sickle cells, and other microbiological markers such as HIV. Furthermore, the oxidative stress and glucose-6-phosphate dehydrogenase (G6PD) statuses of the individuals were unknown. These could affect the levels of the biomarkers reported in this publication. Furthermore, clinical presentations such as anorexia, fever, nausea, and abdominal pain were not investigated in this study. These limitations should be considered in future studies.

Conclusions
Human balantidiosis is endemic in two pig-rearing communities in Ga West Municipality, Ghana, with an overall prevalence of 10.4 %, with direct pig farmers being significantly infected. Most of the people infected with B. coli produced semi-formed, loose and blood stools. Irrespective of the physical appearance of their stool samples, occult blood was present. Individuals infected with B. coli exhibited relative reductions in haemoglobin, red blood cells, platelet count, lymphocytes, chloride, bicarbonate, and albumin, while mean cell volume levels significantly increased. Continuous surveillance for B. coli infection in endemic areas is essential for early treatment to prevent pathologies associated with infected individuals.