New epidemiological features on animal trypanosomiasis by molecular analysis in the pastoral zone of Sideradougou, Burkina Faso

A multidisciplinary work was undertaken in the agropastoral zone of Sidéradougou, Burkina Faso to try to elucidate the key factors determining the presence of tsetse flies. In this study the PCR was used to characterize trypanosomes infecting the vector ( Glossina tachinoides and Glossina palpalis gambiensis ) and the host, i.e. cattle. A 2-year survey involved dissecting 2211 tsetse of the two Glossina species. A total of 298 parasitologically infected tsetse were analysed by PCR. Trypanosoma vivax was the most frequently identified trypanosome followed by the savannah type of T. congolense and, to a lesser extent, the riverine forest type of T. congolense , and by T. brucei . No cases of T. simiae were found. From the 107 identified infections in cattle, the taxa were the same, but T. congolense savannah type was more frequent, whereas T. vivax and T. congolense riverine forest types were found less frequently. A correlation was found between midgut infection rates of tsetse, nonidentified infections and reptile bloodmeals. These rates were higher in G.p. gambiensis , and in the western part of the study area. T. vivax infections were related to cattle bloodmeals, and were more frequent in G. tachinoides and in the eastern study area. The PCR results combined with bloodmeal analysis helped us to establish the relationships between the vector and the host, to assess the trypanosome challenge in the two parts of the area, to elucidate the differences between the two types of T. congolense , and to suspect that most midgut infections were originating from reptilian trypanosomes.     

A phyto-sociological analysis of the distribution of riverine tsetse flies in Burkina Faso

In Burkina Faso, Glossina palpalis gambiensis Vanderplank and G. tachinoides Westwood (Diptera: Glossinidae) are the main cyclic vectors of trypanosomiasis. The vegetation type along river banks is an important factor determining the distribution and abundance of these tsetse. The following work investigated the relation between the plant species present (including the disturbance level) and tsetse distribution and abundance, using three ecotypes, described by P.C. Morel in 1978. These were the Guinean, Sudano-Guinean and Sudanese gallery forests. In the Mouhoun River basin, these three ecotypes are found successively from upstream to downstream. Berlinia grandiflora, Syzygium guineense and Cola laurifolia and finally Acacia seyal and Mitragyna inermis were the best indicators for the Guinean, Sudano-Guinean and Sudanese gallery forest ecotypes, respectively, as suggested by Morel. However, other species such as Pterocarpus santalinoides and Mimosa pigra were not ecotype specific. Trap catches confirmed that G. palpalis and G. tachinoides are predominant in Guinean and Sudanese gallery forests, respectively, and that both species are well represented in the Sudano-Guinean ecotype. Tsetse densities dropped significantly in disturbed Sudano-Guinean and Sudanese gallery forest sites. However, this was not the case for both species in Guinean or for G. tachinoides in half-disturbed Sudanese gallery forest sites, confirming their high resilience to human-made changes. The importance of a detailed consideration of riverine ecotypes when predicting tsetse densities is discussed.     

Population structuring of the tsetse Glossina tachinoides resulting from landscape fragmentation in the Mouhoun River basin,Burkina Faso

The impact of landscape fragmentation resulting from human‐ and climate‐mediated factors on the structure of a population of Glossina tachinoides Westwood (Diptera: Glossinidae) in the Mouhoun River basin, Burkina Faso, was investigated. Allele frequencies at five microsatellite loci were compared in four populations. The average distance between samples was 72 km. The sampling points traversed an ecological cline in terms of rainfall and riverine forest ecotype, along a river loop that enlarged from upstream to downstream. Microsatellite DNA demonstrated no structuring among the groups studied (F_ST = 0.015, P = 0.07), which is contrary to findings pertaining to Glossina palpalis gambiensis Vanderplank in the same geographical area. The populations of G. tachinoides showed complete panmixia (F_IS = 0, P = 0.5 for the whole sample) and no genetic differentiation among populations or global positioning system trap locations. This is in line with the results of dispersal studies which indicated higher diffusion coefficients for G. tachinoides than for G. p. gambiensis. The impact of these findings is discussed within the framework of control campaigns currently promoted by the Pan African Tsetse and Trypanosomosis Eradication Campaign.     

Behavioral interactions and rhythms of activity of Glossina palpalis gambiensis and G. tachinoides (Diptera: Glossinidae) in forest gallery in the Burkina Faso

Glossina palpalis gambiensis and G. tachinoides are the main vectors of human and animal trypanosomoses in West Africa. In some parts of their distribution area, they co-exist in sympatry, but little is known about their interactions. This study aimed to explore their respective flight height and daily activity when co-existing or alone. Attractive targets were used, made of a black/blue/black cloth covered with adhesive film, so that all tsetse that landed were caught. The study was conducted in two areas in South Burkina Faso: Kartasso, upstream the Mouhoun river, where G. p. gambiensis is the only tsetse occurring; and Folonzo, on the Comoé river, where both species occur. Out of more than 3,800 tsetses caught in total, in Folonzo, G. tachinoides occurred at higher densities than G. p. gambiensis (84 vs 16% of the total densities). The mean height of capture was 55 cm for G. tachinoides, and 65 cm for G. p. gambiensis. As a comparison, in Kartasso where G. p. gambiensis is alone, the mean height of capture was 46 cm, these differences being statistically significant. In average, females were caught higher in altitude than males, and the two species showed a similar activity profile in the day. These results are discussed in the light of differences in the nature of the forest gallery, or possible interspecies competition behaviour in relation with their limited energy metabolism and flight capacities, or also with species differences in landing behavior, linked to host feeding detection. These observations have consequences on control tools releasing attractive odours, which may have contrasted efficacy depending of the flight height of the species.     

Intraspecific variability in natural populations of Glossina palpalis gambiensis from West Africa, revealed by genetic and morphometric analyses

Glossina palpalis gambiensis Vanderplank (Diptera: Glossinidae) from West Africa (Senegal and Burkina Faso) were analysed for microsatellite DNA polymorphisms and size of the wings. In the overall sample a strong heterozygote deficiency was found at two polymorphic microsatellite loci. It led to a highly significant value of Fis (within-sample heterozygote deficit) in the western zone of Sideradougou area in Burkina Faso. Genetic differentiation was significant on a macrogeographic scale, i.e. between tsetse coming from Senegal and Burkina Faso. Wing measures also differed between these two countries; flies from Senegal appeared to be smaller. Microsatellite loci further allowed differentiation of populations of G. palpalis gambiensis trapped on the same hydrographic network a few kilometres apart. The results are interpreted as indicating that further investigations will allow the study of genetic variability of tsetse flies in relation to the dynamics of transmission of human and animal trypanosomoses.     

Trypanosoma vivax in Glossina palpalis gambiensis do not appear to affect feeding behaviour, longevity or reproductive performance of the vector

Feeding behaviour of Glossina palpalis gambiensis Vanderplank infected with Trypanosoma vivax Ziemann was studied and compared with that of uninfected control tsetse. The parameters measured were: total number of probes into the ear-skin of rabbits; rate of bloodmeal engorgement; weight of freshly ingested blood; survival; and mean weight of pupae. The results showed that the rosettes of T.vivax parasites in the labrum did not interfere with the feeding behaviour of the vectors. Furthermore, mean survival of T. vivax-infected males was significantly higher (82.2 +/- 4.2 days) compared with that of uninfected ones (70.5 +/- 3.1 days). However, with the female tsetse, mean survival of those infected was lower (98.8 +/- 4.0 days) compared to the uninfected controls (102.2 +/- 5.6 days), but the difference was not significant. A few infected males and females lived a little longer than the uninfected ones. Fecundity of the female tsetse remained unaffected by the infection, and furthermore the mean weight of pupae from the infected females was not significantly different from that of pupae from the uninfected control group. Thus the physiology of pregnant female tsetse in terms of nourishment of intra-uterine larva was unaffected by T.vivax infection. Two successive probes into the skin of two different goats followed by feeding on a third goat by each of four infected tsetse resulted in successful transmission of the infection to eleven out of twelve goats. Thus probing alone into the skin of this host can result in the transmission of T.vivax infection.     

Rate of trypanosome killing by lectins in midguts of different species and strains of Glossina

The activity of lectins in different species of tsetse was compared in vivo by the time taken to remove all trypanosomes from the midgut following an infective feed and in vitro by agglutination tests. Teneral male Glossina pallidipes Austen, G. austeni Newstead and G. p. palpalis R-D. removed 50% of all Trypanosoma brucei rhodesiense Stephens & Fantham infections within 60 h. A 'refractory' line of G. m. morsitans Westwood took 170 h to kill 50% infections while a 'susceptible' line of the same species failed to kill 50%. Agglutination tests with midgut homogenates showed differences between fly stocks which accorded with differences in rate of trypanosome killing in vivo. Flies fed before an infective feed were able to remove trypanosomes from their midguts more quickly than flies infected as tenerals. Increasing the period of starvation before infection increased the susceptibility to trypanosome infection of non-teneral flies. Teneral flies showed little agglutinating activity in vitro, suggesting that lectin is produced in response to the bloodmeal. Feeding flies before infection also abolished the differences in rate of trypanosome killing found between teneral 'susceptible' and 'refractory' G. m. morsitans, suggesting that maternally inherited susceptibility to trypanosome infection is a phenomenon limited to teneral flies. Electron micrographs of midguts of G. m. morsitans suggest that procyclic trypanosomes are killed by cell lysis, presumably the result of membrane damage caused by lectin action.     

Comparative study on the infection rates of different laboratory strains of Glossina species by Trypanosoma congolense

Teneral Glossina morsitans centralis Machado, G.austeni Newstead, G.palpalis palpalis Robineau-Desvoidy, G.p.gambiensis Vanderplank, G.fuscipes fuscipes Newstead, G.tachinoides Westwood and G.brevipalpis Newstead, from laboratory-bred colonies, were fed at the same time on the flanks of ten goats infected with Trypanosoma congolense Broden isolated in Tanzania or in Nigeria. The seven tsetse species were infected over the range 0.3-49.2%. Survival of both T.congolense isolates was best in G.m.centralis, poorest in G.austeni and the four palpalis group tsetse, with G.brevipalpis intermediate. It is suggested that there are differences in the gut of different laboratory-bred cultures of Glossina Westwood species and subspecies such that T.congolense parasites can survive better in the gut of some than in others and undergo cyclical development to metacyclics in the hypopharynx.     

Proteomic Investigation on Anopheles gambiae in Burkina Faso Related to Insecticide Pressures from Different Climatic Regions

In Sub‐Saharan Africa, An. gambiae sensu lato (s.l.) Giles 190, largely contributes to malaria transmission. Therefore, the authors carry out a proteomic analysis to compare its metabolic state, depending on different pesticide pressures by selecting areas with/without cotton crops. The proteomes data are available via ProteomeXchange with identifier PXD016300. From a total of 1.182 identified proteins, 648 are retained for further statistical analysis and are attributed to biological functions, the most important of which being energy metabolism (120 proteins) followed by translation‐biogenesis (74), cytoskeleton (71), stress response (62), biosynthetic process (60), signalling (44), cellular respiration (38), cell redox homeostasis (25), DNA processing (17), pheromone binding (10), protein folding (9), RNA processing (9), other proteins (26) and unknown functions (83). In the Sudano‐Sahelian region, 421 (91.3%) proteins are found in samples from areas both with and without cotton crops. By contrast, in the Sahelian region, only 271 (55.0%) are common to both crop areas, and 233 proteins are up‐regulated from the cotton area. The focus is placed on proteins with putative roles in insecticide resistance, according to literature. This study provides the first whole‐body proteomic characterisation of An. gambiae s.l. in Burkina Faso, as a framework to strengthen vector control strategies.     

Assortative mating in mixed swarms of the mosquito Anopheles gambiae s.s. M and S molecular forms,in Burkina Faso,West Africa

The molecular form composition of Anopheles gambiae Giles s.s. (Diptera: Culicidae) mating swarms and the associated mating pairs (copulae) were investigated during two rainy seasons (July to October, 2005 and July to November, 2006) in the villages of Soumousso and Vallée du Kou (VK7). Although the habitats of these villages differ markedly, sympatric populations of M and S molecular forms of An. gambiae s.s. occur in both places periodically. The main aim was to assess the degree to which these molecular forms mate assortatively. In Soumousso, a wooded savannah habitat, the majority of swarm samples consisted of only S‐form males (21/28), although a few M‐form males were found in mixed M‐ and S‐form swarms. In VK7, a rice growing area, the majority of swarm samples consisted of only M‐form males (38/62), until October and November 2006, when there were nearly as many mixed‐form as single‐form swarms. Overall, ～60% of M‐ and S‐form swarms were temporally or spatially segregated; the two forms were effectively prevented from encountering each other. Of the remaining 40% of swarms, however, only about half were single‐form and the rest were mixed‐form. Of the 33 copulae collected from mixed‐form swarms, only four were mixed‐form pairs, significantly fewer than expected by random pairing between forms (χ² = 10.34, d.f. = 2, P < 0.01). Finally, all specimens of inseminated females were of the same form as the sperm contained within their spermatheca (n = 91), even for the four mixed‐form copulae. These findings indicate that assortative mating occurs within mixed‐form swarms, mediated most probably by close‐range mate recognition cues.     

Polytene chromosome maps in four species of tsetse flies Glossina austeni, G. pallidipes, G. morsitans morsitans and G. m. submorsitans (Diptera: Glossinidae): a comparative analysis

Photographic polytene chromosome maps from pupal trichogen cells of four tsetse species, Glossina austeni, G. pallidipes, G. morsitans morsitans and G. m. submorsitans were constructed and compared. The homology of chromosomal elements between the species was achieved by comparing banding patterns. The telomeric and subtelomeric chromosome regions were found to be identical in all species. The pericentromeric regions were found to be similar in the X chromosome and the left arm of L1 chromosome (L1L) but different in L2 chromosome and the right arm of L1 chromosome (L1R). The L2 chromosome differs by a pericentric inversion that is fixed in the three species, G. pallidipes, G. morsitans morsitans and G. m. submorsitans. Moreover, the two morsitans subspecies appeared to be homosequential and differ only by two paracentric inversions on XL and L2L arm. Although a degree of similarity was observed across the homologous chromosomes in the four species, the relative position of specific chromosome regions was different due to chromosome inversions established during their phylogeny. However, there are regions that show no apparent homology between the species, an observation that may be attributed to the considerable intra—chromosomal rearrangements that have occurred following the species divergence. The results of this comparative analysis support the current phylogenetic relationships of the genus Glossina.     

A comparison of African buffalo, N''Dama and Boran cattle as reservoirs of Trypanosoma congolense for different Glossina species

Teneral Glossina morsitans centralis Machado were fed on the flanks of the African buffalo (Syncerus caffer Sparrman), N'Dama (Bos taurus L.) or Boran (Bos indicus L.) cattle infected with Trypanosoma congolense Broden. The infected tsetse were maintained on rabbits and on day 30 after the infected feed, the surviving tsetse were dissected to determine the infection rates. The mean infection rates (% +/- SE) in the midgut of tsetse fed on buffalo, N'Damas and Borans were 23.5 +/- 3.3, 31.6 +/- 2.7 and 33.7 +/- 4.6, respectively. The differences were not significant. However, the mean mature infection rate in tsetse fed on the buffalo (13.2 +/- 2.1%) was significantly lower compared to the rates in tsetse fed on the N'Dama (20.4 +/- 1.4) or the Boran cattle (21.4 +/- 1.1). When groups of teneral G.m.centralis, G.pallidipes Austen, G.p.gambiensis Vanderplank, G.f.fuscipes Newstead, G.brevipalpis Newstead and G.longipennis Corti were fed simultaneously on either an infected buffalo, an N'Dama or a Boran steer, the mature infection rates ranged from 0 to 16.1%. Irrespective of the host species used, the T.congolense infection rate was highest in G.m.centralis, lowest in the palpalis and fusca group tsetse, with G.pallidipes being intermediate. Nevertheless, the trypanoresistant African buffalo and N'Dama may serve as reservoirs of T.congolense as can trypanosusceptible Boran cattle.     

用PCR直接筛选和鉴定虎纹捕鸟蛛毒素Ⅰ的重组阳性克隆

以合成的两段插入序列为上、下游引物用PCR法直接筛选插入有虎纹捕鸟蛛毒素Ⅰ（HWTX－Ⅰ）cDNA的重组阳性克隆。并用PCR法快速鉴定重组体中插入片段的正、反连接方向,扩增用引物是以位于克隆位点上游的一段载体序列上游引物,以插入序列为下游引物。对100个单克隆进行了上述两次PCR筛选鉴定,选取2个有靶片段插入并且为正向连接的重组子进行测序,其结果证实了插入片段及其方向的正确性。     

Comparative evaluation of carbosulfan- and permethrin-impregnated curtains for preventing house-entry by the malaria vector Anopheles gambiae in Burkina Faso

Pyrethroid-impregnated bednets and curtains are widely employed to reduce the risk of malaria transmission, but pyrethroid-resistance is becoming more prevalent among malaria vector Anopheles mosquitoes (Diptera: Culicidae). As an alternative treatment for curtains, we assessed carbosulfan (a carbamate insecticide) in comparison with permethrin as the standard pyrethroid, against endophilic female mosquitoes of the Anopheles gambiae Giles complex in a village near Ouagadougou, Burkina Faso. The main criterion evaluated was the impact of curtains (hung inside windows, eaves and doorways) on the number of An. gambiae s.l. females active indoors at night. Light-traps were operated overnight (21.00-06.00 hours beside occupied untreated bednets) to sample mosquitoes in houses fitted with net curtains treated with carbosulfan 0.2 g ai/m2 or permethrin 1 g ai/m2 or untreated, compared with houses without curtains. The treated and untreated curtains significantly reduced the numbers of mosquitoes collected indoors, compared with houses without curtains. Carbosulfan-treated curtains had a highly significantly greater effect than permethrin-treated or untreated curtains, the scale of the difference being estimated as three-fold. However, there was no significant difference between the impact of untreated and permethrin-treated curtains on densities of An. gambiae s.l. trapped indoors. Samples of the An. gambiae complex comprised An. arabiensis Patton and both the S- and M-forms of An. gambiae Giles s.s. Susceptibility tests revealed some resistance to DDT and low frequencies of permethrin-resistance, insufficient to explain the poor performance of permethrin on curtains. Among survivors from the diagnostic dosage of permethrin were some specimens of all three members of the An. gambiae complex, but the kdr resistance mechanism was detected only in the S-form of An. gambiae s.s. Questions arising for further investigation include clarification of resistance mechanisms in, and foraging behaviour of, each member of the An. gambiae complex in this situation and the need to decide whether carbosulfan-treated curtains are acceptably safe for use to reduce risks of malaria transmission.     

Identification and differentiation of human schistosomes by polymerase chain reaction

Recent increasing number of travelers, immigrants and foreign workers from schistosomiasis endemic area has thus resulted in the importation of schistosomiasis to non-endemic countries. To avoid ova-induced pathogenicity, sensitive and specific diagnostic means at an early stage of infection are therefore crucial. In this study, we developed polymerase chain reaction (PCR) primers specific for human schistosome species. The PCR products were obtained in a species-specific manner (479 bp, Schistosoma mansoni; 365 bp, S. haematobium; 614 bp, S. japonicum; 303 bp, S. mekongi) and were detectable from 0.01 pg of total worm DNA (S. haematobium, S. japonicum, S. mekongi). The primer sets were also available for multiplex use. Although some difficulties were experienced in amplifying the parasite DNA from the infected animals, schistosome DNA could be detected from one day post infection. The PCR method described herein will therefore be beneficial to detect human schistosomiasis, after some improvements in this method.     

Field application of the polymerase chain reaction (PCR) to the detection and characterization of trypanosomes in Glossina longipalpis (Diptera: Glossinidae) in Côte d'Ivoire

The Polymerase Chain Reaction (PCR) technique was used for the identification of natural trypanosome infections in Glossina longipalpis (Diptera: Glossinidae) in Côte d'Ivoire. A total number of 139 flies were examined microscopically for the presence of trypanosomes. Out of them 50 were detected positive and were subsequently prepared for the PCR using primers specific for Trypanosoma (Nannomonas) congolense of Savannah, Riverine-Forest, Kilifi, and Tsavo types, T. (N.) simiae, T. (Duttonella) vivax and Trypanozoon. Almost 90% of the infections detected by the PCR were attributed to Nannomonas, especially T. congolense Savannah and Riverine-Forest types, with many infections in which both of these two types were present T. simiae and T. vivax were also detected in some flies. The sequence specificity of the PCR products was confirmed by hybridization with parasite-type specific DNA probes. Differences between parasitological and PCR results are discussed.     

Detection of bean yellow mosaic virus in gladioli corms by the polymerase chain reaction

The polymerase chain reaction (PCR) method readily detected bean yellow mosaic virus (BYMV) in gladioli leaves, but in initial tests PCR did not detect virus in corm tissue. Extracts of RN A from corm tissue were shown to inhibit the amplification of viral sequences when added to a PCR reaction. An additional purification step for the RNA extracts using a Sephadex G-50 column eliminated the inhibitory effect and enabled PCR to amplify and detect viral RNA in corm tissue preparations.     

Detection of Plasmodium sporozoites in mosquitoes by polymerase chain reaction and oligonucleotide rDNA probe, without dissection of the salivary glands

Dried Anopheles gambiae mosquito head + thorax portions, infected with Plasmodium falciparum sporozoites, were processed by the polymerase chain reaction. The PCR product was hybridized to an oligonucleotide probe (known as 114R or AW34) diagnostic for Plasmodium . The detection level by autoradiography was ten sporozoites per mosquito. Head + thorax of mosquitoes that contained mature P.falciparum oocysts, without sporozoites, gave no positive signal, indicating that the test detects only infective mosquitoes. This test can be applied to wild mosquito specimens collected, prepared and processed at different time intervals. The technique is convenient, highly sensitive, and could be used with a non-radioactive detection system and specific probes to differentiate Plasmodium spp.     

Detection of Helicobacter pylori DNA in feces and saliva by polymerase chain reaction: a review

The polymerase chain reaction (PCR), known for its high sensitivity and specificity, has been used for the detection of Helicobacter pylori DNA in bodily materials such as feces and saliva. Since fecal specimens contain PCR inhibitors, DNA before PCR amplification has been purified using various biochemical, immunological and physical pre-PCR steps. Several PCR protocols, differing from each other in the selection of genomic targets and primers, have produced varying degrees of specificity and sensitivity in detecting H. pylori DNA. PCR identified antimicrobial resistance of H. pylori in feces. It also detected virulence factor genes such as the cytotoxin-associated gene (cagA) and vacuolating cytotoxin gene (vacA) in feces and saliva. While the cagA gene was detected in 50-60% of fecal specimens, it was found in 25% of salivary specimens from patients. There was considerable variation in the detection rate of H. pylori DNA in salivary samples. The detection rate in saliva with the most effective primer pair was lower than that observed in feces, making saliva a less suitable specimen for the diagnosis of H. pylori infection. There is controversy regarding the permanent presence of H. pylori in saliva. Whether the salivary and gastric specimens of an individual harbor identical or different strains has not been resolved. PCR cannot distinguish between living and dead organisms. However, it can offer quick results on fecal and salivary specimens, which may contain fastidious and slow-growing H. pylori in low numbers.     

Damage formation and repair efficiency in the p53 gene of cell lines and blood lymphocytes assayed by multiplex long quantitative polymerase chain reaction

We examined ultraviolet (UV) irradiation and cisplatin treatment damage formation and repair efficiency in the p53 tumor suppressor gene of various cultured cell lines and lymphocytes using a nonradioactive multiplex long quantitative polymerase chain reaction (QPCR) assay, which amplified a 7-kb fragment of the target gene and a 500-bp fragment of the template control to successfully increase the sensitivity and reliability of the assay. The multiplex long QPCR detected a lesion frequency of 0.63 lesions/10kb/10J/m(2) in the p53 gene of fibroblast cells. In addition, the multiplex long QPCR assay detected pronounced differences in the repair of UV damage in the p53 gene among repair-proficient CRL-1475 cells and repair-deficient XP-A and XP-C cells. The multiplex long QPCR assay was also evaluated as a sensitive assay for the detection of DNA damage induced by cisplatin. The data indicated that the lesion frequency in the p53 gene was 1.27-1.75 times higher in the H23 cisplatin-sensitive cell than in the H1435 cisplatin-resistant cell at the IC(70) dose. After 8-h and 24-h repair periods, only 13 and 75% of cisplatin-induced damage had been removed in the H23 cells, whereas these values were 92 and 100% in the H1435 cells. In addition, our data indicate that multiplex long QPCR is a sensitive method for validly estimating repair in freshly isolated lymphocytes. The results suggest that the current protocol of the multiplex long QPCR method can be used to assess the damage formation and repair efficiency of various agents at biologically relevant doses and to allow a more precise determination of gene-specific repair in disease susceptibility and drug resistance in epidemiological studies.