Individual subject meta-analysis of parameters for Giardia duodenalis shedding in animal experimental models.

Giardia duodenalis is a zoonotic protozoan parasite with public health importance worldwide. While articles about animal model infectivity have been published for G. duodenalis, the studies have used diverse protocols and parameters to evaluate the infectivity of this protozoan parasite. Hence, the objectives of this study were to (1) conduct a meta-analysis of published literature for cyst shedding and diarrhea outcomes in animal models and (2) develop recommendations to help standardize experimental dose response studies. Results showed that, for the outcome of cyst shedding in faeces, the covariates of infective stage (cyst versus trophozoite), Giardia dose, and the interactions between doses and infective stage, as well as dose and species of experimental host, were all significant (P value ≤ 0.05). This study suggests inoculation of the experimental host with cysts rather than trophozoites and administration of higher doses of Giardia will most likely result in cyst shedding. Based on the results of this meta-analysis, the infective stage (cyst versus trophozoite), parasite dose, and the interactions between dose and infective stage, as well as dose and species of experimental host, should be considered when designing experimental dose response studies that will assist in the study of zoonotic neglected tropical diseases globally.

1. Introduction

Giardia duodenalis is a zoonotic waterborne pathogen known to be released into the environment through human and animal faeces [1]. Infection and gastrointestinal disease in humans are caused by ingestion of small doses (>10 cysts) of Giardia cysts [2] directly from faeces or indirectly from contaminated water and food products [1, 3]. Giardia is considered the most common cause of protozoan diarrhea in developing countries and worldwide [1]. It has been estimated that in Africa, Asia, and Latin America about 200 million people have symptomatic giardiasis with some 500000 new cases being reported each year [4]. The number of reported cases in the US was 18,913 in 2008 [5]. In 2004, Giardia was included in the World Health Organization's Neglected Disease Initiative for its link with poverty [6]. G. duodenalis, also called G. intestinalis and G. lamblia, is the only species found in humans [1, 3] and will be referred to as G. duodenalis throughout this paper.

Animal and human dose-response models as well as in vitro cell culture have been widely used to evaluate Giardia spp. infectivity. At present, human infectivity studies are not considered to be practical because of the ethical concerns involved. Therefore, very few experiments have involved the attempted infection of humans with Giardia isolates [2, 7]. Instead, experimental animal models have been frequently used to estimate whether G. duodenalis is strictly host specific or zoonotic [1] and to provide information used for human health risk assessments. In vitro cell culture has been proposed as a potential alternative to animal assays because cultivation has the advantage of being cheaper and less time consuming and does not raise the ethical concerns associated with using animal models [8]. However, experimental animal model studies are often preferred when performing risk assessment, as they provide the most realistic scenario of what is happening with the pathogen once it enters the host. For this reason, this meta-analysis focuses on the diverse range of experimental animal model studies available in the peer-reviewed literature.

Numerous cross-transmission experiments involving the infection with isolates of Giardia in a variety of animal species have been published. Each of the experimental studies used different parameters such as animal species, age, cysts or trophozoites, dose, and detection methods to quantify cysts in faeces and evaluate the infectivity of G. duodenalis as well as outcomes ranging from laboratory detection of cyst shedding to clinical symptoms. Using diverse variables and protocols can be problematic in experimental studies because the variations in protocols between studies can affect the outcome of the experiments and thus their comparability [1, 9]. Consequently, a standardization of the experimental dose-response studies is required.

Hence the main objective of the present study was to develop recommendations for the standardization of animal dose-response experiments by conducting a meta-analysis of individual subjects and exploring study design characteristics that cause heterogeneity between included studies. The hypothesis tested in this meta-analysis was that selected experimental factors of interest are associated with increased G. duodenalis cyst shedding and diarrhoea. The results of this meta-analysis will be useful for more standardized and comparable studies of Giardia infectivity, so that in the future there may be fewer animal species and resources used to provide dose-response information more quickly and efficiently.

2. Methods

2.1. Literature Search Strategy and Selection of Studies

The search of published animal and/or human dose-response studies in all languages was performed using the electronic databases PubMed and Web of Science from March to December, 2010. The process is described in Figure 1, with search criteria consisting of the following algorithms: (i) protozoan infectious doses for humans, (ii) Giardia and dose response, (iii) Giardia and infectivity, (iv) Giardia and meta-analysis, (v) giardiasis and experimental and model, and (vi) experimental and infection and Giardia. Unpublished studies (grey literature) were not included in this meta-analysis.

Figure 1

Flow of information through the different phases of a systematic literature review for the Giardia duodenalis meta-analysis.

The articles identified by means of these search criteria were subject to a further selection process consisting of the removal of the complete study or individuals within the articles that met any of the exclusion criteria (Table 1). One reviewer examined the titles and abstracts of all the articles found using the search criteria mentioned above. The full article for each of the relevant studies was then assessed by the same reviewer and a second reviewer was consulted as needed. Only those articles that met at least one of the search criteria and none of the exclusion criteria were included in the data analysis. Furthermore, the reference lists of all included articles were searched for further possible papers, but no additions were identified. Ethical approval was not required for this meta-analysis.

Table 1

Exclusion criteria for studies and number of articles excluded from the Giardia duodenalis meta-analysis.

Exclusion criteria	Articles excluded
(1) Giardia  different than G. duodenalis, G. intestinalis,or G. lamblia were used	19
(2) Infection was assessed only by histology on tissue sections	6
(3) Number of animals with diarrhoea or shedding was not reported	4
(4) Animal study groups were subject to treatments other than the single inoculation/single outcome design used in this meta-analysis	3
(5) Infection was assessed by cyst shedding and histology together and it was not possible to distinguish how many animals were shedding	4
(6) Experimental dose response was measured in humans only	2
(7) Study was not an experimental dose response experiment	1
(8) Cyst dose administered to the experimental host was not provided or a range was provided	1
(9) Number of animals used was not provided	1

2.2. Data Extraction

Data regarding the following variables for each individual animal were extracted and recorded from each article: (i) Giardia species (Giardia), (ii) isolate analyzed, such as HP-10 (Isolate), (iii) assemblage of the Giardia sp. such as A (Assemblage), (iv) original source of isolate (Isolate Source), (v) whether cysts or trophozoites were subject to passage before challenge (Passage), (vi) storage time of cysts or trophozoites (in weeks) prior to inoculation of experimental host (Storage Time), (vii) method used to confirm cyst or trophozoite viability prior to inoculation (Viability Method), (viii) animal species used as experimental host (Experimental Species), (ix) number of subjects per group inoculated (Number), (x) age of the experimental host (Age), (xi) whether the experimental host was subject to immunosuppression (Immunosuppression), (xii) method of immunosuppression of the experimental host (Immunosuppression Cause), (xiii) whether cysts or trophozoites were administered to the experimental host (Infective Stage), (xiv) cyst or trophozoite dose administered to the experimental host (Giardia Dose), (xv) administration route used to inoculate the cysts or trophozoite dose to the experimental host (Administration Route), (xvi) cyst detection method in faeces of experimental hosts after inoculation (Detection Method), and (xvii) the number of animals that shed cysts and/or presented with diarrhoea after inoculation.

The unit of analysis was the individual animal; thus, the value for each variable was collected for each animal and included in the analysis. If a variable was not reported at the animal level, the variable was left blank. Table 2 provides information regarding the number of subjects and missing values by variable and study. The primary outcomes of interest were the presence of cyst shedding and diarrhoea. An animal was considered to have shedding or diarrhoea when the condition of the animal was described with the words diarrhoea or cyst shedding in the study from which the information was extracted. All of the studies included in this analysis evaluated shedding over prolonged periods so even if shedding was intermittent, any report of cysts observed in the feces was considered as positive for shedding. For diarrhoea and cyst shedding, the classification of Yes/No was used. Initially, attempts were made to contact authors for clarification, but, due to lack of responses, this approach was not systematically implemented throughout the entire study.

Table 2

Studies included for Giardia duodenalis meta-analysis of diarrhoea (D) and shedding (S) outcomes (chronological order).

Reference	Year	Outcome	Sample size (number of animals)
			IS1	P2	ST3	ES4	A5	IS6	GD7	AR8	DM9
[23]	1978	S10	20	20	20	20	20	20	20	20	20
[24]	1979	S10	47	47	NR11	47	47	47	47	47	NR11
[17]	1982	S10	11	11	11	11	11	11	11	11	NR11
[19]	1982	S10	150	150	150	150	150	150	150	150	150
[25]	1984	S10	49	49	NR11	49	49	49	49	49	NR11
[26]	1985	S10	25	25	14	25	NR11	25	25	25	25
[18]	1986	S10	60	60	NR11	60	49	60	60	60	60
[27]	1988	S10	204	204	107	204	204	204	204	204	204
[28]	1989	S10	10	10	NR11	10	NR11	10	10	10	10
[29]	1990	S10	75	75	NR11	75	75	75	75	75	75
[30]	1991	S10	NR11	NR11	NR11	10	10	10	10	10	10
[31]	1991	S10	109	109	109	109	109	109	109	109	109
[32]	1992	S10	NR11	14	NR11	14	14	14	14	14	14
[33]	1993	S10	62	62	NR11	62	62	62	62	62	62
[34]	1994	S10	4	16	NR11	16	16	16	16	NR11	16
[35]	1996	S10	10	10	10	10	10	10	10	10	10
[36]	1997	S10	NR11	10	NR11	10	10	10	10	10	10
[37]	2002	S10	NR11	94	NR11	94	94	94	94	94	94
[38]	2005	S10	56	56	NR11	56	56	56	56	56	56
[39]	2006	S10	NR11	6	NR11	6	6	6	6	6	6
[40]	2007	S10	97	97	97	97	97	97	97	97	97
[41]	2007	S10	6	6	NR11	6	6	6	6	6	6
[42]	2008	S10	NR11	12	NR11	12	12	12	12	12	12
[43]	2008	S10	NR11	60	NR11	60	60	60	60	60	60
[44]	2010	S10	40	40	NR11	40	40	40	40	40	40
[45]	2010	S10	33	33	33	33	33	33	33	33	33
[46]	1995	S and D12	7013, 2814	7013, 2814	3013, 814	7013, 2814	7013, 2814	7013, 2814	7013, 2814	4013, 2014	7013, NA15
[47]	1995	S and D12	5213, 3014	5213, 3014	2213, NR11	5213, 3014	5213, 3014	5213, 3014	5213, 3014	5213, 3014	5213, NA15
[48]	1997	S and D12	1016	1016	NR11	1016	1016	1016	1016	1016	NA15
[49]	2010	S and D12	1416	1416	NR11	1416	1416	1416	1416	1416	NA15

1Isolate Source, 2Passage, 3Storage Time (weeks), 4Experimental Species, 5Age, 6Infective Stage, 7 Giardia Dose, 8Administration Route, 9Detection Method, 10Shedding outcome, 11Information was not reported in study; thus no individual data were included, 12Shedding and Diarrhoea outcome, 13Number of individuals for shedding outcome, 14Number of individuals for diarrhoea outcome, 15Does not apply, 16Number of individuals for both outcomes.

2.3. Classification of Variables

The extracted variables were classified according to the information provided by the selected studies for each outcome (Table 3). In the case of the variable “Isolate Source,” the classification “Other than Humans” was composed of ruminants (cattle and lambs), primates, rodents, and drinking water. For the variable “Passages,” the classification “Yes” was composed of in vivo (animal passage), in vitro (cultures), and both in vivo and in vitro passages, while the classification “No” corresponded to cysts that were not subject to any passage. For the “Storage Time” variable, the class “<1 week” included animals that were inoculated with cysts or trophozoites that were stored for less than one week; “≥1 week” included all those animals that were inoculated with cysts or trophozoites that were stored for one week or more. Treating the variable “Storage Time” as a continuous variable, with or without log transformation, did not improve the model fit. Regarding the variable “Experimental Species,” the category “Other Animals” included dogs, cats, and cattle, while the class “Other Rodents” comprised rats and hamsters. The variable “Age” was categorized following published criteria (Table 4), which were consistent with the criteria used in the studies included in this meta-analysis. The classification of age was dependent on the animal model as animal species mature at different rates. For instance, 2-month-old kittens are in a different maturity stage than 2-month-old gerbils and thus might have different susceptibilities to protozoan infection. When ranges of age were reported, the mean of the range was used in the meta-analysis. The classification “Young” in the “Age” variable for the shedding outcome consisted of newborn and weanling animals, while the classification “Adult” included only adult animals. For the diarrhoea outcome, no adult animals with diarrhoea were reported. Therefore, the “Age” variable for the diarrhoea outcome was composed of “Newborn” and “Weanling.” For the variable “Administration Route,” “Gastric Intubation” included the terms gastric intubation, intragastric route, and stomach tube, while “Other Than Gastric Intubation” included the terms orally, intraesophageal route, and nasogastric. The variable “Detection Method” was only included in the cyst shedding outcome and included three classes: “Hemocytometer,” “Microscopy,” and “Flotation Techniques.” “Microscopy” included indirect fluorescent antibody (IFA), fluorescent microscopy, and Nomarski interference contrast (NIC) microscopy. “Flotation Technique” included parasite concentration methods such as zinc sulfate flotation and sucrose gradient flotation. The variable “Giardia Dose” was a continuous variable. As the values in the “Giardia Dose” variable ranged from 4 to 107 cysts or trophozoites, the data for this variable were log transformed to make the distribution of the data more normal. The mean of the whole dose series (mean log scale = 4.12) was subtracted from every dose value to center the data. This enabled interpretation of the main effect in the presence of an interaction term at a meaningful value (the mean) [10]. For instance, in the multivariable model, the main effect of “Infective Stage” in the model that also includes the interaction “Giardia Dose by Infective Stage” corresponded to a difference between cysts and trophozoites at a mean dose. If the mean of the whole dose series was not subtracted, this comparison would have been the difference between cysts and trophozoites at a log dose of zero (dose of 1).

Table 3

Classification of categorical variables for both Giardia duodenalis cyst shedding and diarrhoea outcomes.

Variable	Shedding outcome			Diarrhoea outcome
	Classification of variables1	Number of studies	Number of individuals	Classification of variables1	Number of Studies	Number of individuals
Assemblage	A	3	148	A and E	1	14
	B	1	49
	E	1	6
	A and E	1	14
Isolate Source	HumansOther Than Humans	194	112886	HumansOther Than Humans	13	2854
Passage	NoYes	1716	755667	NoYes	22	2458
Storage Time	<1 week≥1 week	103	420183	<1 week	1	8
Experimental Species	MiceOther AnimalsOther RodentsGerbils	78415	18087202963	Other AnimalsGerbils	22	2458
Age	AdultYoung	1218	458828	NewbornWeanling	22	2458
Infective Stage	CystsTrophozoites	2018	1021411	CystsTrophozoites	32	6418
Administration Route	Gastric IntubationOther Than Gastric Intubation	623	2331153	Gastric IntubationOther Than Gastric Intubation	22	5024
Detection Method	HemocytometerMicroscopyFlotation Technique	1539	74953523	Not analyzed

1Categories not listed indicate that there were no animals from those categories in the studies.

Table 4

Criteria used to determine age categories for experimental species (alphabetical order).

Animal species	Criteria used
	Newborn	Weanling1	Adult
Cattle	<21 weeks	21–64 weeks [50]	≥65 weeks [51]
Cats	<8 weeks	8–25 weeks [52]	≥26 weeks [51]
Dogs	<8 weeks	8–29 weeks [53]	≥30 weeks [51]
Gerbils	<4 weeks	4–7 weeks [54]	≥8 weeks [55, 56]
Hamsters	<3 weeks	3 weeks [57]	>3 weeks
Mice	<3 weeks	3 weeks [58]	>3 weeks [51]
Rabbits	<4 weeks	4 weeks–17 weeks [59, 60]	≥18 weeks [51]
Rat	3 weeks	3–7 weeks [61]	≥8 weeks[51]

1Weanling was merged with the newborn category to create the classification “Young” in the “Age” variable for the cyst shedding outcome.

2.4. Statistical Analyses

In this meta-analysis, the statistical analysis and publication bias assessment were done as described in Adell et al. [11]. Final models were selected based on inclusion of the four key variables “Experimental Species,” “Age,” “Giardia Dose,” and “Infective Stage,” allowing for as many interaction terms as possible and having a combination of variables that generated narrower confidence intervals or more precision. The fit of the model was assessed by the ratio of the generalized chi-square statistic and its degrees of freedom (generalized chi-sq/df), which could not exceed a value of 1, and residual plots. The statistical software JMP 9 (SAS 2010) and SAS 9.3 (SAS 2011) were used to perform all the analyses.

3. Results

3.1. Search Strategy and Selection of Studies

As shown in Figure 1, the initial search identified 1045 potentially relevant studies. Subsequent to reviewing the abstracts, 71 studies were considered for further screening, of which 41 were not eligible since at least one of the exclusion criteria (Table 1) was met. After examination of the full text, 30 studies were included for the cyst shedding outcome while 4 were incorporated for the diarrhoea outcome. The studies included in this meta-analysis were published between 1978 and 2010 and are shown in Table 2. The total number of individuals included in this meta-analysis was 1432 individuals for the shedding outcome and 82 individuals for the diarrhoea outcome.

3.2. Analysis for the Cyst Shedding Outcome

The bivariate analysis showed that the covariates “Passage,” “Experimental Species,” “Age,” “Infective Stage,” and “Giardia Dose” were associated with cyst shedding in faeces (P value ≤ 0.2) (Table 5). These variables were then incorporated in the multivariable analysis which provided one model that best fulfilled the selection criteria. The final multivariable model (Table 6) shows that the variables “Infective Stage” and “Giardia Dose” and the interactions “Giardia Dose by Experimental Species” and “Giardia Dose by Infective Stage” have at least one category with a statistically significant difference (P value ≤ 0.05) from the reference category (Table 6). The variables “Experimental Species,” “Age,” and “Administration Route” were not significant, but they were incorporated into the model to control for confounding and effect modification. “Administration route” and “Age” were identified as a potential confounder (an epidemiologic term specifically describing a variable associated with a change in the coefficient estimate of at least one variable when placed in the model) of the relationship between “Giardia dose” and cyst shedding and “Experimental Species” and cysts shedding, respectively. “Experimental species” was part of the interactions “Giardia Dose by Experimental Species.”

Table 5

Bivariate analysis results for risk factors associated with Giardia cyst shedding outcomes.

Variable1	P-value	Statistical association between variable and cyst shedding (Yes/No)2
Passage	<0.00013	Yes
Storage Time	0.92693	No
Experimental Species	<0.00013	Yes
Age	0.00233	Yes
Infective Stage	<0.00013	Yes
Giardia Dose	0.00614	Yes

1Data sparseness prevented analytic calculations for Isolate Source, Administration Route and Detection Method.

2Significant level ≤0.2.

3Mantel-Haenzel bivariate analyses analysis taking into account the correlation between studies.

4GLIMMIX bivariate analyses.

Table 6

Multivariable generalized linear mixed model showing risk factor associations for Giardia duodenalis cyst shedding outcomes1.

Variable	Categories	Odds ratio estimate	95% Wald confidence limits	P value
Experimental Species	Gerbils2
	Mice	0.163	(0.003, 8.87)	0.37
	Other Animals	0.034	(<0.001, 5.05)	0.18
	Other Rodents	4.305	(0.10, 183.05)	0.45
Age	Young2
	Adult	0.36	(0.05, 2.43)	0.30
Infective Stage	Trophozoites2
	Cysts	5.026	(2.63, 9.56)	<0.00112
Giardia Dose	Not applied (continuous variable)	2.677	(1.81, 3.94)	<0.00112
Administration Route	Gastric Intubation2
	Other Than Gastric Intubation	10.39	(0.19, 573.61)	0.25
Giardia Dose by Experimental Species	Mice	4.368	(1.59, 11.93)	0.00412
	Other Animals	2.129	(0.12, 37.45)	0.61
	Other Rodents	0.0910	(0.01, 0.72)	0.0212
Giardia Dose by Infective Stage	Cysts	0.5711	(0.41, 0.81)	0.00212

1Pseudo-AIC = 7013.49.

1Generalized chi-square statistic and its degrees of freedom (gener. Chi-sq/df) = 0.66.

2Corresponds to the reference category.

3The odds ratio estimate corresponds to Mice versus Gerbils at the mean log dose.

4The odds ratio estimate corresponds to Other Animals versus Gerbils at the mean log dose.

5The odds ratio estimate corresponds to Other Rodents versus Gerbils at the mean log dose.

6The odds ratio estimate corresponds to Cysts versus Trophozoites at the mean log dose.

7The odds ratio estimate corresponds to an increase of one in log dose in Gerbils with Trophozoites.

8The odds ratio corresponding to an increase of one in log dose in Mice is 4.36 times that in Gerbils.

9The odds ratio corresponding to an increase of one in log dose in Other Animals is 2.12 times that in Gerbils.

10The odds ratio corresponding to an increase of one in log dose in Other Rodents is 0.09 times that in Gerbils.

11The odds ratio corresponding to an increase of one in log dose in Cysts is 0.57 times that in Trophozoites.

12Risk factors with statistically significant results.

The multivariable model indicated that inoculating cysts into the experimental host had 5.02 times higher odds of cyst shedding than inoculating trophozoites at the mean log dose of 0.0538 (P value ≤ 0.001; CI: 2.63, 9.56). For each 1 unit of change in the log of trophozoite dose administered to the reference experimental host “Gerbil,” the odds of cyst shedding increased 2.67 times (P value < 0.0001; CI: 1.81, 3.94), whilst the odds ratio corresponding to an increase of one in log dose in cysts was 0.57 times that in trophozoites (P value = 0.002; CI: 0.41, 0.81).

The model also indicated that, for each unit of increase in the log dose administered to “Mice,” the odds of cyst shedding increased 4.36 times compared to “Gerbils” (P value: 0.004; CI: 1.59, 11.93). Whereas, for each unit increase in the log dose administered to “Other Rodents,” the odds of having cyst shedding were reduced 0.09 times compared to “Gerbils” (P value: 0.02; CI: 0.01, 0.72). Figure 2(a) shows the range of doses compared to the odds of detecting shedding by experimental species, illustrating how the choice of the experimental animal species does not have an effect until beyond a log scale 6, with mice only having an increase in odds. When “Mice” was removed from the analysis, the odds of shedding cysts were similar for all animals until log 7 after which the odds of shedding were greater in “Gerbils” as opposed to “Other animals.” Figure 2(b) shows the ranges of Giardia doses compared to the odds of detecting shedding and how choosing young experimental animals had a higher impact on shedding in response to different doses being investigated compared to adult experimental animal.

Figure 2

Range of Giardia dosages captured in the articles included in this meta-analysis by the odds of the cyst shedding outcome for the categories “Age” and “Experimental Species.” (a) Range of Giardia dosages (log scale) by odds of cyst shedding by Experimental Species. (b) Range of Giardia dosages (log scale) by odds of cyst shedding by Age.

The fit of the model was assessed by the ratio value for generalized chi-sq/df value and residual plots. The generalized chi-sq/df for the multivariable model was 0.66, indicating that the model had good fit. The residual plots (not shown), in which the residual values were plotted against the linear predictor, showed that two observations in each model had large residuals (≥15) and were possible outliers from the cluster of observations. Removing the individuals that had large residuals did not change the significance of the other variables of the model; thus, they were kept in the model.

In this meta-analysis, we found evidence of possible publication bias for the cyst shedding outcome. The funnel plot showed a higher number of studies over the mean value (Figure 3). The Egger regression test suggested a significant association between study size and study effect (P value of 0.002). The Duval and Tweedie trim and fill method suggested adding 10 studies to the left side of the funnel plot, which under the random effects model would shift the point estimate of the prevalence of cyst shedding of all the studies included in this meta-analysis from 0.77 (CI: 0.68, 0.85) to 0.61 (CI: 0.49, 0.72), improving the estimate. The results of Cochran's Q test indicated significant heterogeneous results (P value < 0.001) among different studies and the I 2 statistic determined that 88% of variation across studies was due to significant heterogeneity rather than random chance.

Figure 3

Funnel plot of standard error by logit event rate for cyst shedding outcome after Giardia infection showing possible publication bias as evidenced by a higher concentration of studies to the right of the mean.

3.3. Analysis for the Diarrhoea Outcome

As the data for the diarrhoea outcome were sparse, the analysis done for the shedding outcome could not be performed for the diarrhoea outcome. Therefore, the bivariate analysis was done considering the data as a pool across studies to estimate associations rather than accounting for the correlation between studies as done for the shedding outcome. The bivariate analysis showed that, for the outcome of diarrhoea in exposed animals, the covariates “Isolate Source” and “Experimental Species” were significant (P value ≤ 0.1) (Table 7). Due to the smaller number of studies reporting on the outcome of diarrhoea and variability in the reporting of these potential covariates across studies, it was not possible to create a multivariable model for this outcome. The publication bias for diarrhoea outcome was not assessed due to the low number of studies (4 studies) included in this meta-analysis.

Table 7

Bivariate analysis results for risk factors associated with diarrhoea outcome after Giardia infection.

Variable1	P value	Statistical association between variable and presence of diarrhoea (Yes/No)2
Isolate Source	<0.00013	Yes
Passage	0.353	No
Experimental Species	0.073	Yes
Age	0.353	No
Infective Stage	0.253	No
Giardia Dose	0.254	No
Administration Route	0.323	No

1Data sparseness prevented analytic calculations for Storage Time.

2Significant level ≤0.2.

3Mantel-Haenzel bivariate analyses not taking into account the correlation between studies.

4GLIMMIX bivariate analyses.

4. Discussion

This meta-analysis evaluated the effects of multiple experimental covariates on cyst shedding or diarrhoea as indicators of G. duodenalis infection. The results obtained in this meta-analysis identified covariates that potentially cause heterogeneity between the outcomes of the dose results experiments and suggest that administering cysts or trophozoites to experimental hosts and the dose administered can all significantly impact the incidence of cyst shedding.

Based on the results of this meta-analysis, it would be appropriate to make the following recommendations for future dose-response experiments on G. duodenalis when assessing infection by means of the presence of cyst shedding in experimentally infected animals to make the studies more comparable and increase likelihood of cyst shedding: (i) use cysts to infect the experimental animals rather than trophozoites, (ii) consider the infective stage used (cysts versus trophozoites) and the Giardia dose administered together, (iii) consider the animal species used as an experimental host and dose together, and (iv) taking this into account, consider using mice as experimental hosts rather than gerbils, rats, hamsters, dogs, cats, and cattle. These parameters should be considered when designing experimental dose-response studies, as once the designs of the dose-response studies are more standardized, they will provide better information and more comparable results for more accurate risk assessments that consider infection as the outcome.

In the case of assessing infection by means of the presence of diarrhoea in experimentally infected animals, more experimental studies in animal models should be conducted, as not enough studies have been reported to obtain estimates of the effect of different experimental parameters on diarrhoea in individual animals. It should be noted that while the diarrhoea outcome is of clinical relevance, the presence of asymptomatic infected individuals is a limitation of using diarrhoea to represent infectious status. Based on the bivariate analyses, it would be appropriate to consider and report the following in future dose-response experiments: (i) the assemblage being inoculated into the experimental host, (ii) original source of the cysts being inoculated, (iii) whether cysts or trophozoites were subject to any passage before inoculation into the experimental host or not, (iv) animal species used as experimental hosts, (v) age of the experimental host, and (vi) the administration route used to inoculate the cysts or trophozoites into the experimental host. These studies will provide better information and more comparable results for risk assessments that consider illness as an outcome.

Interestingly, the significant “Giardia Dose by Experimental Species” interaction suggests that an increase of one in log of dose administered has a larger impact on cyst shedding in “Mice” than “Gerbils” as experimental hosts, indicating a larger difference between the experimental host-groups as the administered dose increases. But, an increase of one in log of dose administered had a protective impact on cyst shedding in “Other Rodents” compared to “Gerbils” as experimental hosts. Nevertheless, differences in cyst shedding among experimental animal hosts may depend on the dose of Giardia administered, and it is advisable to consider these two variables together when designing experimental studies. In addition, it is important to take into consideration that some species might vary in their susceptibilities to different G. duodenalis assemblages. For instance, dogs are the only species that have been reported to be susceptible to G. duodenalis assemblage C and D, cats to assemblage F, hoofed livestock to assemblage E, and rats to assemblage G [1, 3, 12], while a wide variety of animals, such as cattle, dogs, cats, rodents, and other wild animals and humans are susceptible to G. duodenalis assemblage A [1, 3, 12]. Thus, choosing to use mice, other rodents, or other animals as experimental hosts would be appropriate only for some study objectives, such as evaluating the infectivity of the different G. duodenalis assemblagesby means of cyst shedding.

Pathogen shedding patterns for newborns, weanlings, and adults can be quite different across host species. For instance, it has been reported that weanling calves generally lack a strong specific humoral immune response to G. duodenalis infection, while newborns may be protected by the anti-Giardia activity of colostrums [13], indicating that weanlings may be more susceptible to Giardia infection than newborns or adults. Unfortunately, in this meta-analysis there were not enough studies to analyse newborn and weanling animals separately, thus both categories had to be merged and analyzed as “Young.” Based on the meta-analysis results, young animals were not statistically different from adult animals with regard to cyst shedding. This finding is in contrast to numerous studies that reported the prevalence and cyst excretion peaks in young animals [14-16]. However, our results from this meta-analysis are in concordance with the findings reported by Hewlett et al. [17], where young mongrel dogs were not found to be more susceptible than adults, and Woo and Paterson [18] where adult and young dogs and cats did not present with infection after being challenged with 300000 cysts. The data available for this meta-analysis were unfortunately inadequate to provide estimates to determine how the dose and the choice of the age of the experimental species impact the odds of cyst shedding.

The multivariable model suggested that inoculating the experimental host with cysts would increase the likelihood of cyst shedding compared to inoculation with trophozoites. However, the model also indicated that an increase of one log in dose had less of an impact on cysts than it did on trophozoites. Therefore, differences by using cysts versus trophozoites for host infection depend on the Giardia dose used, and it is advisable to consider these two variables together when designing experimental studies. Results of the multivariable model also suggest that when a higher dose of trophozoites is administered to gerbils, the odds of cyst shedding increase. Numerous studies in which different doses of Giardia cysts or trophozoites have been administered to animal models have been published. However, whether the infectious dose may contribute to symptom variability is still unclear [1]. Studies have shown that the infectivity ratio is directly related with the number of cysts inoculated to the experimental host [19, 20]. Nevertheless, all of these reports were individual studies; therefore, there is the possibility that those results were influenced by the experimental design or other factors. Among the strengths of meta-analyses is that it provides an overall estimate of an association or effect based on a number of independent scientific studies and explores the variation in the observed effect across studies, thus obtaining a gain in statistical power to detect effects [21]. The result obtained in the multivariable model indicated that when a higher dose is administered, the odds of having cyst shedding increases.

Information for the variables “Assemblage,” “Isolate Source,” and “Storage Time” was scarce and thus the association between these variables and the shedding outcome could not be assessed. Passage of cysts or trophozoites used to infect experimental hosts was identified as a risk factor for cyst shedding in the bivariate analysis. However, information on this variable was not reported for all studies, which made it challenging to consider it jointly in a multivariable model. Once more experimental studies providing information regarding these variables are published, it would be beneficial to evaluate the association with the shedding outcome and incorporate them in the multivariable model “Experimental Species,” “Age,” “Infective Stage,” and “Giardia Dose” to ascertain the independent effect of these possible covariates on the presence of cyst shedding.

This meta-analysis showed evidence of possible publication bias for the cyst shedding outcome. This finding can be explained by the fact that smaller studies are more likely to be published if they have larger than average effects, which makes them more likely to meet the criterion for statistical significance [22]. The addition of studies by the trim and fill procedure improved the point estimate of the prevalence of cyst shedding of all the studies included in this meta-analysis. However, the adjusted estimate is similar to the original effect and thus indicates that the reported trends may be valid. Caution is advised when interpreting these results because there is evidence of high heterogeneity among the studies, thus precluding a full evaluation of publication. These findings suggest that it is important to publish those studies that have negative or nonsignificant results as well those that have significant or positive results in order to reduce or avoid the bias.

Based on the results of this meta-analysis, it is crucial that more experimental studies in animal models are conducted to assess infectivity of Giardia by means of cysts shedding and diarrhoea. These studies would provide useful data for risk assessments that consider either infection or illness as an outcome.

5. Conclusions

When assessing G. duodenalis infection using a cyst shedding outcome measure, this study suggests that differences among the animal species used as experimental hosts depend on the dose of Giardia administered. It is therefore advisable to consider these two variables (host and dose) together when designing experimental studies. Taking this into account, mice appear to be the most appropriate animal model in which to assess infection when using a cyst shedding outcome, as they were more likely to shed cysts than other animal species.

Young and adult animals were similarly likely to shed cysts. Nevertheless, additional studies are needed for increased statistical power to ascertain effects of the log dose increment for different age groups on the presence of cyst shedding.

For considering whether cysts or trophozoites are used to challenge the experimental hosts to assess infection by means of cyst shedding, the multivariable analysis results suggest that it would be more appropriate to use cysts. However, it also indicated that an increase of one log dose has less of an impact for cysts than for trophozoites; thus, differences between inoculating cysts or trophozoites in the cyst shedding depend on the dose administrated. As expected, administering higher doses of cysts or trophozoites increases the odds of cyst shedding.

When using a diarrhoea outcome measure in experimental studies, the source of the isolate and species of experimental animal host should be considered when designing experimental studies. As additional studies are published, greater power will be possible to distinguish individual and joint effects of the identified covariates.