Utility of lingual examination in identifying sentinel pigs for porcine cysticercosis serological studies in Taenia solium endemic areas

H A Ngowi

Department of Veterinary Medicine and Public Health, Sokoine University of Agriculture, P. O. Box 3021, Morogoro, Tanzania
helenangowi@gmail.com

Abstract

Lingual examination is currently the method of identifying pigs for experimental studies on Taenia solium cysticercosis in endemic areas.  The utility of lingual examination in identifying sentinel pigs intended for monitoring seroconversion to T. solium larval antigen was assessed. 

Sensitivity of lingual examination in detecting positive results of antigen enzyme-linked immunosorbent assay (Ag-ELISA) was 31.3% in 251 pigs aged 4-7 months and 24.1% in 440 pigs aged 9-12 months.  In both age groups the specificity was 99%.  The likelihood ratio for a positive lingual examination result was 31.3 and 24.1 in pigs aged 4-7 and 9-12 months, respectively.  The likelihood ratio for a negative lingual examination result was 0.7 and 0.8, respectively.  Lingual examination of 2-3 months old pigs in three different studies had predictive values for a negative test of 94.9% (95% CI: 93.2, 96.7; n = 609), 84.1% (95% CI: 77.2, 91.0; n = 107) and 100% (95% CI: 100, 100; n = 96).  Lingual examination is a suitable method for identifying pigs for monitoring Ag-ELISA seroconversion.  The method is more suitable for identifying infected than cysticercosis-free pigs as it suffers from a high likelihood of false negative results.  Lingual examination is not suitable for monitoring the incidence of porcine cysticercosis seroconversion nor can it enable approval of pig meat for human consumption.  Studies are needed to determine the utility of lingual examination in selecting pigs intended for monitoring T. solium antibody seroconversion.  Rapid antigen detection tests are needed to complement with the lingual examination to reduce the likelihoods of infected pigs from entering the food chain in endemic areas.     

Key words: likelihood ratios, seroprevalence, tongue cysts

Introduction

Porcine cysticercosis is caused by Taenia solium, a parasite that infects humans (definitive hosts) and a few intermediate hosts, humans inclusive.  In most endemic areas, pigs serve as the most common intermediate hosts that transmit taeniasis (adult parasite infection) as a result of consumption of infected pork.  Given the short life span, pigs are considered important sentinels for monitoring environmental contamination with T. solium eggs (Gonzalez et al 1994), which can highlight on the risk of human infection with cysticercosis (infection with the larval form).  In addition, infection rates in pigs can be used to monitor efficacy and effectiveness of T. solium control strategies. 

Pigs have been used globally in experimental studies assessing efficacy or effectiveness of T. solium control strategies, including health education (Sarti et al 1997; Ngowi et al 2008), treatment of infected pigs (Mkupasi et al 2013) and vaccination.  While in porcine cysticercosis-free countries sentinel pigs can easily be identified for experimental studies, in endemic countries, care is needed to correctly identify non-infected pigs for studies that require pigs free from cysticercosis at the beginning of the study.  Unfortunately, there are currently no accurate and quick diagnostic tests for identifying sentinel pigs in endemic countries.  Serological studies that aim at detection of circulating antigens for cysticercosis such as those using antigen enzyme-linked immunosorbent assay (Ag-ELISA) would need to start with pigs that are free from cysticercosis, and hence, circulating antigens to allow detection of new infections. 

Most antigen detection assays have high sensitivity and specificity in detecting porcine cysticercosis.  For example, Dorny et al (2004) reported sensitivity and specificity of 86.7% and 94.7%, respectively.  Nevertheless, such assays are not currently readily available, especially at the time of enrolment of pigs.  Even if available, such assays would require a couple of days before obtain the results, especially for large studies.  This would require capturing the pigs for sampling and release them until the assay results are out before a decision to recruit them is made.  With the dynamic nature of pig populations in rural areas, this is impractical in such areas.

While efforts in developing quick diagnostic tests are underway, examination of the undersurface of pig’s tongue (lingual examination) remains the currently practical and available method for initial screening of pigs for cysticercosis.  This method has been estimated to have a sensitivity of 21% (95% CI: 14–26%) and specificity of 100% (Dorny et al 2004) in detecting porcine cysticercosis.  These values may vary depending on the raters and the statuses of the lingual cysts (viable or calcified).  It has been common in many experimental studies in porcine cysticercosis endemic countries that pigs are recruited in a trial based on their cysticercosis status as revealed by lingual examination (Ngowi et al 2008, Mkupasi et al 2013).  Such pigs are then followed up with serological tests (antibody or antigen detection) alone or in combination with other tests.  For the detection of viable infections in pigs, Ag-ELISA has been the most common assay employed.  The current analysis determined the accuracy of lingual examination in indicating Ag-ELISA actual results for accurate recruitment of pigs intended for porcine cysticercosis experimental studies that employ Ag-ELISA.  Specifically, the analysis assessed the sensitivity, specificity and likelihood ratios of the lingual examination in comparison with the Ag-ELISA.  This study also assessed how the lingual examination was able to correctly identify Ag-ELISA negative piglets in three different studies.   

Materials and methods

Study area

This study involved secondary analysis of data collected during three previous field randomised trials to determine effectiveness of interventions to reduce incidence rate of porcine cysticercosis in Mbulu district, northern Tanzania (Ngowi et al 2008; Mwidunda 2012).  Mbulu district is situated in the northern highlands of Tanzania, between latitudes 3.80^o and 4.50^o S and longitudes 35.00^o and 36.00^o E.  The altitude ranges from 1000 to 2400 above sea level.  The rainfall ranges from <400 mm to >1200 mm.  The major economic activity in the district is crop and livestock production. 

Data

The first dataset was collected during a pre-post randomised field trial to evaluate the effect of health education on knowledge, attitudes, practices and incidence rate of porcine cysticercosis in Mbulu district, northern Tanzania, between 2002 and 2004 (Ngowi et al 2008).  A total of 733 piglets aged 2-3 months from cysticercosis-free sows based on lingual examination were introduced to smallholder pig farmers (one per farmer) having been confirmed free from cysticercosis also based on lingual examination (Ngowi et al 2008) .  The blood of the pigs was sampled at the time of recruitment for later confirmation of cysticercosis status using antigen enzyme-linked immunosorbent assay (Ag-ELISA).  These pigs were re-examined using Ag-ELISA and lingual when they were 4-7 and 9-12 months old.  The flow of the pigs followed by both lingual and Ag-ELISA for the current analysis is presented in Figure 1.  The second dataset was collected in a field randomised trial to estimate the effect of TSOL18 vaccine in the prevention of porcine cysticercosis in Mbulu district, northern Tanzania in 2012.  A total of 120 piglets aged 2-3 months free from cysticercosis based on lingual examination had their blood sampled for later confirmation of their cysticercosis infection status based on Ag-ELISA (Mwidunda 2012 Tanzania Food and Drugs Authority, personal communication).  The third dataset was collected in a second field randomised trial to further confirm the effect of TSOL18 vaccine in the prevention of porcine cysticercosis in Mbulu district, northern Tanzania in 2014  (Mwidunda 2014 Tanzania Food and Drugs Authority, personal communication).  In this study, 140 piglets aged 2-3 months free from cysticercosis based on lingual examination, had their blood collected for later serological confirmation.  The flow of the pigs in the three studies is presented in Figure 2.

Data analysis

Utility of the lingual examination method in the screening of pigs intended for serological studies based on antigen detection method was assessed through analysis of the likelihoods of the lingual examination method in indicating the serological status of the pigs aged 4-7 and 9-12 months that were introduced in cysticercosis endemic area.  Thereafter, the lingual examination method was assessed on its ability to predict Ag-ELISA-free pigs that were introduced in the two vaccination trials.

Analysis of likelihoods of lingual examination in indicating Ag-ELISA status of the pigs aged 4-7 and 9-12 months

First, a two-by-two table was drawn for each age category, with the lingual examination results against those of Ag-ELISA.  Sensitivity and specificity of the lingual examination method to indicate the Ag-ELISA status were calculated.  The sensitivity and specificity were used to calculate the likelihood ratios for positive (LR+) and negative (LR-) lingual examination tests.  The likelihood ratios were calculated as (Akobeng 2006):

LR+ = Sensitivity/(1-Specificity)

LR- = (1-Sensitivity)/Specificity

The guidelines presented in Appendix 1 were used in the interpretation of the likelihood ratios in this study.

Analysis of the probability of lingual examination method to correctly identify negative Ag-ELISA results

The negative predictive values (NPV-) and their 95% confidence intervals for the lingual examination method were calculated for each of the 609, 107 and 96 two to three months old piglets that tested negative based on lingual examination.  For the 609 piglets which came from lingual cyst-free m sows, a Chi-square test was performed to analyse for any association between sow’s Ag-ELISA positivity with that of their offspring.  This information would help to realise as to whether double screening (screening sows followed by screening their piglets) was superior to single screening (screening piglets only) in identifying sentinel piglets.

Results

Sensitivity, specificity and likelihood ratios of lingual examination in indicating Ag-ELISA status in 4-7 and 9-12 months old pigs

The sensitivities of the lingual examination in detecting Ag-ELISA positive cases were 31.3% and 24.1% in pigs aged 4-7 and 9-12 months, respectively.  The specificity was 99.0% in both categories of pigs (Table 1).  The LR+ and LR- are presented in Table 1.  There LR+ values indicated large and often conclusive increase in the likelihood of Ag-ELISA positivity when the lingual examination results were positive.  On the other hand, the LR- values indicated minimal decrease in the likelihood of Ag-ELISA negativity when the lingual exam results were negative. 

Predictive values of lingual examination method

The NPV- of lingual examination against Ag-ELISA ranged from 84.1% to 100% in the three studies (Table 2).  As expected, these values increased with decrease in prevalence of the infection.

Association between piglet Ag-ELISA seropositivity and that of its mother

There was not any significant association between piglet seropositivity and that of their mothers (Chi-square = 0.2986, P = 0.5848, n = 609).

Discussion

This study has for the first time reported the utility of lingual examination in identifying sentinel pigs for experimental studies employing detection of circulating antigens of T. solium larvae.  The sensitivity of lingual examination in detecting Ag-ELISA positive pigs was below 40%, with some variation between pig age groups.  This variation is not surprising considering that the Ag-ELISA is not a gold standard test for cysticercosis.  It is possibly that the sensitivity of Ag-ELISA also varied between the two studies.  Nevertheless, the low sensitivity of lingual examination to Ag-ELISA makes it unsuitable for monitoring infection incidence as it cannot confirm Ag-ELISA negativity. 

The present analysis estimated a high specificity of lingual examination in detecting Ag-ELISA negative results.  The specificity was constant in the two pig age categories.  Because of the study design, it was not possible to determine sensitivity and specificity of lingual examination to Ag-ELISA in pig 2-3 months old, the age at which many intervention trials recruit pigs.  Nevertheless, as these parameters are expected to remain stable, the author would anticipate similar values in any age of the pigs and infection prevalence. The apparent 1% loss of specificity of lingual examination to Ag-ELISA can be speculated on the possibility of the Ag-ELISA detecting Taenia hydatigena larvae, which could also infect pigs.  This cannot be detected by lingual examination.  However, the observed high specificity of lingual examination to Ag-ELISA makes the lingual examination a suitable test for confirming positive Ag-ELISA serological status with high probability.  Thus the method is more suitable for identifying infected than cysticercosis-free pigs as it suffers from a high probability of false negative results.     

Likelihood ratios were used in this analysis as stable parameters to determine the relationship of lingual examination results with the Ag-ELISA results.  This analysis found that, a pig with a positive Ag-ELISA result was about 24 to 31 times more likely to have a positive lingual examination result than a pig that had a negative Ag-ELISA.  On the other hand the analysis revealed that the probability of having a negative lingual examination result for pigs with positive Ag-ELISA results was 0.7 to 0.8 times that of those with negative Ag-ELISA results, a minimal decrease in the likelihood.  In other words, pigs with negative Ag-ELISA results were about 1.3 to 1.4 times more likely to have a negative lingual examination result than pigs with positive Ag-ELISA results.  Based on the observed high LH+, lingual examination is a good test for rejecting most of the pigs that have circulating antigens of T. solium.  Nevertheless, the observed low likelihood for a negative lingual examination results indicates that a certain percentage of pigs with circulating antigens for T. solium may be missed by the lingual examination method.  This could be partly due to the Ag-ELISA detecting T. hydatigen, which could also be infecting some pigs or just because of the low sensitivity of the lingual examination method.  Researchers intending to use the lingual examination method to select Ag-ELISA free pigs for studies should account for a potential false negative proportion by increasing the sample for the study to allow omission of this proportion of pigs during data analysis without reducing the power of the study.  Blood samples should be collected at the time of recruitment of pigs to enable later identify those pigs whose Ag-ELISA positivity could not be detected by the lingual examination at the time of recruitment to the study.  One of the best ways to reduce the likelihood of false negative pigs from entering studies or pork market is employing a quick antigen detection test, which should be used in series with the lingual examination to optimise cost effectiveness.  This test is yet to be established although efforts are underway.  Such screening test could be applied in pigs testing negative in lingual examination to increase the rate of detecting lingual false negative pigs.   

The present study found high predictive values for negative tongue examination method in 2-3 months old piglets in three different studies.  These values though high, they decreased with increased prevalence of circulating antigens.   This was expected scientifically as predictive values are not stable parameters of a diagnostic test as compared to likelihood ratios.  Surprisingly, double screening did not seem to have any added advantage in increasing the negative predictive values.  This was also supported by the observed lack of any significant association between the piglet's serological status and that of the sow.  This suggests that screening only piglets should be sufficient in identifying a high proportion of sentinel piglets.

In many rural areas, pig traders use lingual examination to reject pigs if they have T. solium cysts under their tongues.  While the method can help the trader to minimise the chances of buying infected pigs, most of infected pigs remain undetected by the method.  Careful postmortem inspection of pork is required to further reduce the chances of consumption of infected pork.  Nevertheless, the postmortem has low sensitivity, similar to that of lingual examination (Dorny et al 2004).  This calls for thorough cooking of pork in areas where porcine cysticercosis is endemic.  Long-term solutions should include development of rapid and accurate methods of detecting infected pigs before slaughter to safeguard consumers and avoid economic losses.

One of the limitations of this study is the exclusion of 2-3 month old pigs that were positive for cysticercosis based on lingual examination.  This made it impossible to determine sensitivity, specificity and likelihood ratios of lingual examination to Ag-ELISA in pigs aged 2-3 months.  Although these parameters should ideally be constant for a test if the comparison test is gold standard, in the current study they could vary given that the Ag-ELISA is not gold standard in detecting porcine cysticercosis.  Another limitation of this study is the use of predictive values in estimating probabilities of lingual examination in predicting porcine cysticercosis Ag-ELISA seronegativity in pigs aged 2-3 months.  Being unstable parameters, the observed values cannot be generalised as they could change based on disease prevalence.  The PV- would be higher in areas with lower prevalence and lower in areas with higher prevalence of porcine cysticercosis than the current study areas.  Future studies should collect necessary data to enable computation of likelihood ratios in this pig age category.  This study analysed the ability of lingual examination in predicting circulating antigens of T. solium in pigs.  Nevertheless, some studies are based on antibody rather than antigen detection.  Further studies are needed to determine the ability of the lingual examination in predicting T. solium antibody status of pigs intended for antibody seroconversion studies. 

Acknowledgements

The author acknowledge the Danish International Development Agency (DANIDA) and the Association for Strengthening Agricultural Research in Eastern and Central Africa (ASARECA) for funding the previous studies used in the current analysis.

References

Gonzalez A E, Gilman R, Garcia H H, McDonald J, Kacena K, Tsang V C, Pilcher J B, Suare, F, Gavidia C and Miranda E 1994 Use of sentinel pigs to monitor environmental Taenia solium contamination. The Cysticercosis Working Group in Peru (CWG). American Journal of Tropical Medicine and Hygiene 51: 847–850.

Sarti  E, Flisser A, Schantz P M, Gleize, M, Loya M, Plancarte A, Avila G, Allan J, Craig P, Bronfman M and Wijeyaratne P 1997 Development and evaluation of health education intervention against Taenia solium in a rural community in Mexico.  American Journal of Tropical Medicine and Hygiene 56: 127–132.

Ngowi H A, Carabin H, Kassuku A A, Mlozi M R S, Mlangwa J E D. and Willingham III A L 2008 A health education intervention trial to reduce porcine cysticercosis in Mbulu District, Tanzania.  Preventive Veterinary Medicine 85: 52–67.

Dorny P, Phiri I K, Vercruysse J, Gabriel S, Willingham III A L, Brandt, J, Victor B, Speybroeck N and Berkvens D 2004 A Bayesian approach for estimating values for prevalence and diagnostic test characteristics of porcine cysticercosis.  International Journal of Parasitology 34: 569–576.

Akobeng A K 2006 Understanding diagnostic tests 2: likelihood ratios, pre- and post-test probabilities and their use in clinical practice.  Acta Paediatrica 96: 487-491.

Mkupasi E M, Ngowi H A, Sikasunge C S, Leifssona P S and Johansen M V 2013 Efficacy of ivermectin and oxfendazole against Taenia solium cysticercosis and other parasitoses in naturally infected pigs.  Acta Trop., 128(1):48-53.

Received 18 May 2015; Accepted 31 May 2015; Published 2 January 2016

Go to top