Veterinary Microbiology, 30 (1992) 151-163 Elsevier Science Publishers B.V., Amsterdam
151
Comparison of diagnostic procedures for porcine leptospirosis R.J. Chappel a, R.W. Prime b, B.D. Millarb, L.J. Mead a, R.T.
Jones b and
B. A d l e r c
aDepartment of Agriculture, Victorian Institute of A nimal Science, 475 Mickleham Road, A ttwood, Victoria 3049, Australia hDepartment of Agriculture, Regional Veterinary Laboratory, Bendigo, Victoria 3550, Australia CDepartment of Microbiology, Monash University, Clayton, Victoria 3168, Australia (Accepted 5 August 1991 )
ABSTRACT Chappel, R.J., Prime, R.W., Millar, B.D., Mead, L.J., Jones, R.T. and Adler, B., 1992, Comparison of diagnostic procedures for porcine leptospirosis. Vet. Microbiol. 30:151-163. Kidneys and matched serum samples were obtained from 368 pigs slaughtered at three Victorian abattoirs, and originating from 42 farms. Macroscopic lesions (white spots) were observed on 102 of the kidneys. Serum samples were tested by the microscopic agglutination test (MAT) and by an IgM enzyme immunoassay (EIA). Kidneys were cultured for leptospires, examined histologically after Warthin-Starry silver staining and after immunogold silver staining (IGSS), and tested for leptospiral DNA by DNA hybridization. Forty-four infected pigs were identified by culture or immunogold silver staining of kidneys or by high MAT titres (~> 1024). Infection was demonstrated in 7.5% of visibly normal kidneys, in 23.5% of kidneys with white spots, and in 48% of kidneys with large white spots, of I cm diameter or greater. The apparent (maximum) sensitivities of diagnostic procedures for detecting infection were as follows: MAT (at a titre of either 64 or 1024) 95%; IgM EIA 82%; culture 61%; presence of white spots 55%; IGSS 52%; presence of large white spots 30%; WarthinStarry silver staining 20%. IGSS, Warthin-Starry staining and DNA hybridization all appeared to be highly specific. Of 22 kidney sections identified as positive by IGSS, 13 showed intact leptospires, and these kidneys were all culture-positive. Nine others showed leptospiral antigen in the kidney tubules but no intact leptospires. Only five of these kidneys were culture-positive.
INTRODUCTION
Leptospirosis is a significant disease of pigs throughout the world. Leptospira interrogans serovar pomona is the most important serovar affecting pigs, and is responsible for abortions and stillbirths following the infection of susceptible gilts and sows. Such reproductive losses can be controlled by vaccinating breeding females with killed leptospiral vaccines. However leptospirosis can be maintained in growing pigs destined for slaughter, despite the vaccination of the sows from which they are derived. Infected grower pigs 0378-1135/92/$05.00 © 1992 Elsevier Science Publishers B.V. All rights reserved.
152
R.J. CHAPPEL ET AL.
present a h u m a n infection hazard at the abattoir, and meat inspectors and abatoir workers often contract leptospirosis from slaughtered pigs. Infected grower pigs are asymptomatic and cannot usually be identified when they are inspected before slaughter. Unless laboratory tests are performed, the only indication that a slaughtered pig may be infected with leptospires is the presence in some such animals of visible multifocal white lesions in the renal parenchyma ("white spots") (Jones et al., 1987; Hunter et al., 1987; Baker et al., 1989). These lesions are a reflection of interstitial nephritis, which is a c o m m o n result of kidney infection by leptospires. It is important to evaluate the specificity and sensitivity of laboratory diagnostic procedures. Furthermore, because white spots are commonly assumed to indicate leptospirosis, their value as a diagnostic indicator is also of interest. The purpose of this work was twofold. The primary aim was to obtain information on the diagnostic value and proper interpretation of several techniques used for the diagnosis of porcine leptospirosis. The second aim was to obtain information about the relationship between leptospiral infection and white spots observed at slaughter. MATERIALS AND METHODS
Collection and processing of kidneys and blood samples A single kidney with a paired blood sample was collected from each of 368 pigs at three Victorian abattoirs (at Altona, Castlemaine and Daylesford) between March 1988 and March 1989. Blood was collected at the point of slaughter, and the pig was marked with an individual tattoo which was subsequently used to identify the carcass when the kidney was collected at another part of the abattoir. The kidneys were not a random sample but were designed to provide a representative cross-section of kidneys with white spots and visibly normal kidneys. Collection was thus deliberately biased towards kidneys with white spots. No more than 10 kidneys with white spots and 10 visibly normal kidneys were accepted from any one farm.
Grading of kidneys for macroscopic lesions Kidneys were considered to be positive for white spots if they had at least two visible lesions. Kidneys with white spots were classified as having large lesions (about 1 cm diameter or greater) or having small lesions (generally pinhead size or smaller). Individual kidneys which had both large and small lesions were classified as having large lesions. All kidneys were graded by one veterinary pathologist.
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSP1ROSIS
15 3
Culture Kidneys were brought to the laboratory and cultures were set up as soon as practicable, always on the day of collection. A piece of each kidney about 1 cm 3, containing both cortex and medulla, was obtained aseptically from the inside of the kidney, and homogenized in a stomacher in 3 ml of sterile EMJH base. Three drops of the resulting homogenate were added to 5 ml of EMJH m e d i u m (Turner, 1970), supplemented with sodium pyruvate and sodium acetate (each 0.1 g/l) and with added 5-fluorouracil and actidione (each 0.1 g/l), and 3-drop serial transfers were then made into two further 5 ml volumes of medium. Cultures were incubated at 30 ° C, and the contents examined by dark field microscopy after 1-2 weeks, 6 weeks, 10 weeks and approximately 5 months. Kidneys were excluded from the study if all three cultures became contaminated by the 10 week inspection.
Histology Sections of the same piece of kidney, fixed in 10% buffered neutral formalin and processed in paraffin, were used for Warthin-Starry silver staining and for immunogold silver staining. Silver staining by the Warthin-Starry technique (Warthin and Starry, 1920) was performed as modified by Young ( 1969 ). One veterinary pathologist examined all slides under code. Immunogold silver staining (IGSS) of sections of formalin-fixed paraffinembedded kidney was performed as previously described (Skilbeck and Chappel, 1987). Slides were stained by one worker but were examined independently by two workers. Negative and positive control kidneys were included in each batch of slides processed.
DNA hybridization DNA hybridization ( D N A H ) , using a genomic probe derived from serovar pomona, was performed by a modification of the method of McCormick et al. (1989). Approximately 2 g of kidney tissue was homogenized in 2 ml of phosphate-buffered saline (PBS), pH 7.3 and samples were centrifuged at 800 g for 15 min. Equal volumes ( 500/4) of the clear supernatant and 0.8 M NaOH containing 20 m M EDTA were mixed and boiled for 10 min. To 500 fll of the supernatant after boiling, 50/zl of 3 M sodium acetate were added and the solution mixed. One volume of cold 95% ethanol was then added and the tube was shaken. Sample tubes were centrifuged for 5 min in an Eppendorf centrifuge, and the supernatant was removed twice using a vacuum pump. Samples were then resuspended in 10/A of TE buffer and heated at 60 ° C for 15 min, then a 3/tl sample was spotted onto a nylon filter had been prewetted with 6 X SSC and dried. The filter was then placed on blotting paper wet with 6 X SSC and exposed to u.v. light for 2 min. Hybridization was then performed as described by McCormick et al. (1989).
154
R.J. CHAPPEL ET AL.
Sixty-four of the kidneys were tested. Of these, 30 had infection proven by culture or by IGSS. Of the remainder, 22 were from herds from which the kidneys sampled had no proven infection and no white spots, and from which all sera were MAT-negative. The last 12 were from herds in which some of the pigs sampled had positive MAT titres a n d / o r white spots.
Microscopic agglutination test The microscopic agglutination test (MAT) (Faine, 1982 ) was performed using viable serovar pomona as antigen, on sera at final dilutions from 1/4 to 1/8192. Enzyme immunoassay An enzyme immunoassay for IgM antileptospiral antibodies (IgM EIA) was modified from that described by Ballard et al. (1984). Leptospires of serovar pomona at a concentration of 3 X 107/m| were disrupted by sonication and autoclaved to produce the EIA antigen. The wells of U-bottomed polystyrene plates (Nunc Polysorp ) were coated for 16-40 h at 4 oC with 100 ~tl volumes of EIA antigen. 100 ~1 of serum diluted in PBS-Tween containing 1 m g / m l bovine gamma globulin, and 100 #1 of peroxidase-conjugated antiporcine IgM (/~-chain specific) (Kirkegaard & Perry Laboratories) were diluted in PBS-Tween and were consecutively incubated at 37°C for 1 h. Substrate solution (100/zl) consisting of 1 m M 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS) in 0.10 M citrate-phosphate buffer pH 4.2, containing 0.009% hydrogen peroxide was added for 1 h at room temperature and the absorbance was read at 414 nm. Sera were tested in duplicate and were considered positive if the mean absorbances exceeded the mean absorbance of eight replicates of a standard serum containing 64 units of arbitrarily-defined antibody activity. A serum sample obtained from a young pig ten days after experimental infection with L. interrogans serovar pomona was defined as having 1024 units of activity. This sample was rich in IgM antibody, and was diluted in preinfection serum to obtain the 64 unit standard. The m i n i m u m diagnostic value of 64 units was adopted after testing sera from four experimentally-infected pigs (Ballard et al., 1984) using standard curves, and this value provided a compromise between false positive and false negative diagnostic reactions in those animals. All positive results were confirmed by retesting in another assay. Statistical analysis The analysis of ordinal categorical data was done by logistic regression using cumulative logits (McCullogh, 1980). Rating agreement between tests was measured by Cohen's Kappa test (Fleiss, 1973 ), in which K< 0 indicates less than chance agreement, K> 0 indicates more than chance agreement, and K= 1 indicates perfect agreement.
155
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSP1ROSIS
RESULTS
Kidneys and sera collected Kidneys and sera were obtained at three abattoirs from 42 farms, with from 1 to 20 kidneys coming from each individual farm (mean 8.8; median 10). A total of 102 kidneys with white spots was obtained, and these were derived from 28 of the farms. Of the 102 kidneys with white spots, 27 had large lesions ( 1 cm or greater), and these were derived from 13 farms.
Definition of an infected pig After analysis of the results presented below, individual pigs were considered to be infected if leptospires were demonstrated in the kidney by culture of IGSS, or if they had MAT titres I> 1024. A herd was considered to be infected if it contained at least one infected animal.
Culture A total of 27 out of the 368 kidneys was culture-positive (Table 1 ). Of these 27, leptospires were first seen after 1-2 weeks in 11, after 6 weeks in 15, and after 10 weeks in one. No additional positives were detected at the 5 months examination. There was no consistent pattern as to which dilution or dilutions became positive first, but in some cases all three dilutions had visible leptospires after 1-2 weeks. The culture-positive kidneys were derived from five farms of the 42. All were among the 28 farms from which kidneys with white spots were obtained, and four of the five were among the thirteen herds from which kidneys with large white spots were obtained. TABLE 1 R e s u l t s o b t a i n e d ~ o m 44 p i g s c l a s s i f i e d as p o s i t i v e b y c u l t u r e , i m m u n o g o l d s i l v e r s t a i n i n g , o r M A T 1024 Culture
IGSS
MAT
EIA
17
+
+
) 1024
+
12 5 2 2 2 1 1
+ + + -
+ + + -
) ) ) ) ) ) ~
+ + + -
1
-
+
1
+
-
No.
o f pigs
1024 1024 1024 1024 1024 1024 1024 32 0
-
156
R.J. CHAPPEL ET AL.
Immunogold silver staining A total of 21 kidneys was found to be positive by both workers, and one additional positive was detected by each, giving 23 IGSS-positive kidneys in all (Table 1 ). There was good agreement between the two workers (K= 0.84, P / = 1024. 2A herd was regarded as infected if it was shown to have one or more pigs regarded as infected. TABLE 3
Relationship between the results of the MAT and of the IgM EIA in individual sera IgM EIA
Negative Positive
Numbers of Sera MAT negative
MAT 4-256
MAT >/1024
195 73
39 19
6 36
White spots Table 4 compares the diagnostic results obtained with pigs with large white spots, with white spots of any size, or with visibly normal kidneys. Whereas white spots in general, and large white spots in particular, were clearly associated with leptospiral infection, white spots were not a particularly good predictor of current infection ( x = 0.18 ). Only 23.5% of pigs with kidneys with white spots were shown to be infected, and 7.5% of pigs with visibly normal kidneys were also shown to be infected. Large white spots were a better predictor of infection than white spots in general, and 48% of all pig kidneys with these spots were shown to be infected. Likewise 44% of pigs with large white spots on their kidneys had serum MAT titres of 1024 or above, a serological result shown above to be strongly associated with current infection. Among the 44 infected pigs, white spots were present on the kidneys of only 24 (55%), and large white spots on the kidneys of 13 (30%). A total of 14 of the 42 herds sampled in this study yielded no kidneys with white spots, and 116 pigs were sampled from these 14 herds. None of this group of pigs was
158
R.J. CHAPPEL ET AL.
TABLE 4 Diagnostic results obtained with pigs whose kidneys had large white spots, had white spots of any size. or were visibly normal Diagnostic result Total number Infected 3 Culture-positive IGSS positive total intact leptospires antigen only Warthin-Starry positive MAT>~4 M A T ) 1024 EIA positive
Large white spots 27 13 (48%) 8 (30%) 9 (33%) 3 (11%) 5 (19%) 3 (11%) 14 (52%) 12 (44%) 17 (63%)
Any white spots 102 24 (24%) 16 (16%) 16 (16%) 8 (8%) 7 (7%) 6 38 23 46
(6%) (37%) (23%) (45%)
No white spots 266 20 (8%) 11 (4%) 7 (3%) 5 (2%) 2 (1%) 3 62 19 82
(1%) (23%) (7%) (31%)
All pigs 368 44 (12%) 27 (7%) 23 (6%) 13 (4%) 9 (2%) 9 100 42 128
(2%) (27%) (11%) (35%)
3pigs w e r e considered to be infected if their kidneys were culture-positive or positive to IGSS, or if they had MAT titres > / = 1024.
shown to be infected. However 18 of them (16%) had serum antibody detectable by MAT (titres from 4 to 256), and 31 (27%) were positive by IgM EIA. DISCUSSION
Of the 368 pigs studied, 44 (12%) were considered to be infected at the time of slaughter. These included 27 whose kidneys were culture-positive, four whose kidneys were culture-negative but positive by IGSS, and a further 13 which had MAT titres >1 1024. It is possible that additional pigs were infected at the time of slaughter but were not readily detected by either culture or IGSS or by MAT, because other studies have shown the presence of infection in the kidneys of serologically negative pigs (Sturdza et al., 1969; Whyte and Ratcliff, 1982; Whyte et al., 1982). The studied population was drawn from 42 farms but was not randomly selected, as there was a deliberate attempt to collect kidneys with white spots. The proportion of kidneys with white spots that pass through Victorian abattoirs is not known, but is probably less than the 28% (102 out of 368) included in this study. Approximately 14% of pig kidneys inspected in South Australian abattoirs have white spots (P.R. Davies, personal communication). In this study, 24 of the 102 kidneys with white spots (23.5%) had proven infection, as against 20 of the 266 visibly normal kidneys (7.5%). The latter figure provides a lower limit on the prevalence of infection in the population studied. If 10% of all kidneys passing through the abattoirs studied are assumed to have white spots, the minimum level of infection can be esti-
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSPIROSIS
| 59
mated as 7.5 × 0.90 + 23.5 X 0.10 = 9%. Thus 9% or more of pigs slaughtered at these Victorian abattoirs during the study period were probably infected. A proportion of the animals studied had probably been infected during their lifetime, but kidney infection had ceased by the time of slaughter. The duration of shedding of leptospires in the urine of infected pigs is highly variable and may extend from a few days up to a year or more, but a few weeks may be most usual (Ryley and Simmons, 1954; Ryley, 1956; Mitchell et al., 1966; Edwards and Daines, 1979; Chappel et al., 1985 ). Thus a pig that is infected before three months of age may no longer be infectious at slaughter at about five months of age. It can nevertheless be expected to have a significant MAT titre at slaughter (Ballard et al., 1984), and some MAT-positive pigs in this study may therefore have been infected in the past. In this study 100 pigs had serum MAT titres from 4 to 8192 or above. If we discount the lowest group of titres, from 4 to 16, to eliminate any suggestion that the titres did not represent specific antibody attributable to infection, 65 titres of 32 and above remain to indicate that 18% of the pigs studied were infected or had been infected during their lifetime. Of these 65 pigs from 17 herds, 43 were infected, another three were pigs in proven infected herds, and 19 were from herds not known to be infected.
The value of different diagnostic procedures Evaluation of sensitivity requires the identification of a known positive group of animals, such as the 44 identified in this study (Table 1 ). Table 5 shows estimates of the apparent sensitivities (Martin, 1977) of the procedures used. In this study it was established that false negative diagnoses can be given by culture, IGSS, the Warthin-Starry technique, DNAH and IgM TABLE5 Apparent sensitivity of diagnostic procedures for detecting pigs proven to be infected in this study. These figures may be overestimates, as some infected pigs may not have been identified Diagnostic Procedure
Number of Infected Pigs Positive
Apparent Sensitivity
MAT (/>4) MAT (>/64) MAT ( >~1024) IgM EIA Culture White spots IGSS Large white spots Warthin-Starry DNAH
43 42 42 36 27 24 23 13 9 (7/30)
98% 95% 95% 82% 61% 55% 52% 30% 20% (23%)
160
R.J. CHAPPEL ET AL.
EIA. Other studies have shown that the MAT can also be negative in pigs infected with serovar p o m o n a (Sturdza et al., 1969; Whyte and Ratcliff, 1982; Whyte et al., 1982 ), and in this study one infected pig was serologically negative. It was therefore not possible to identify a group of pigs that were uninfected from the population studied, in order to derive numerical estimates of specificity. It was however apparent that the specificities of the Warthin-Starry technique, IGSS and DNAH were all excellent. Culturing detected 61% of known infected pigs over a period up to ten weeks. By definition it does not give false positive diagnostic results. IGSS detected 52% of known infected pigs. It appeared to give no false positive results. In 14 of 22 cases leptospires seen were visibly intact on IGSS. In each case the kidney was also culture-positive. In the other nine cases, of which only five were culture-positive, black dots were visible in the tubules, although absent from negative controls. Scanziani et al. (1989) suggested that such lesions represent later stages of infection, in which disintegration both of leptospires and of the tubules themselves can be observed. The IGSS was shown to be more sensitive than Warthin-Starry silver staining, and of high specificity. It has a role to play in the diagnostic laboratory as a more rapid alternative to culture. The well-established Warthin-Starry technique, although specific, was of low sensitivity in this study, detecting only 20% of infected animals. In our hands IGSS was a more satisfactory alternative. Hunter et al. (1987) failed to detect leptospires by the Warthin-Starry technique in any of 20 culturepositive kidneys. DNAH using a genomic probe derived from serovar pomona, originally applied successfully to urine (Millar et al., 1987), was of low sensitivity but specific when applied to kidney tissue. The sensitivity figure obtained is not directly comparable with the others, being drawn from a subgroup of the infected pigs. The results obtained are in agreement with those of McCormick et al. (1989), who found the technique to be of similar sensitivity to Warthin-Starry staining, but much less sensitive than culture. The culture-positive kidneys studied by McCormick et al. were among those collected in this study, but the technique used was not identical. DNAH has a number of general advantages in leptospirosis diagnosis. Like IGSS, it is much faster than culture, but unlike IGSS it can be readily applied to the simultaneous assay of large groups of samples. Although low sensitivity was observed in this study, adoption of the technique of polymerase chain reaction technology as reported for the diagnosis of bovine leptospirosis (Van Eys et al., 1989) may make DNAH the most sensitive technique available. The MAT was of high sensitivity. Although MAT-negative infected pigs have been reported in the literature, only one such animal was found in this study. Almost all infected pigs had high MAT titres. The significance of low MAT titres (256 or less) is not completely clear. Some of the pigs concerned may have been infected in the past, or may have been in early infection before
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSPIROSIS
161
kidney infection was established. Many low titres (4 to 256) were observed in sera from herds not proven to be infected (Table 2 ), and it is possible that infection was present in these herds. When this work was started it was anticipated that, as antibodies detected by IgM EIA were shorter-lived in the sera of experimentally-infected pigs than were MAT titres (Ballard et al., 1984), the IgM EIA would give a more specific indication of current infection that the MAT. It was therefore expected that fewer sera would be positive by the IgM EIA than to the MAT, but this proved not to be the case. The work of Ballard et al. (1984) was performed in pigs experimentally-infected at eight weeks of age, which is several weeks earlier than pigs commonly become infected in the field in Victoria and South Australia (P.R. Davies and R.W. Prime, unpublished observations). From our unpublished field observations, infection at a greater age seems to produce a more prolonged IgM antibody response. The relationship between the results of the IgM EIA and those of the MAT (Table 3 ) was poor (~c= 0.26, P < 0.001 ). The MAT detects agglutinating surface epitopes, predominantly on the lipopolysaccharide, whereas EIA also detects antibodies against subsurface epitopes. The IgM EIA was less sensitive than the MAT in detecting infected pigs, and there was no evidence that it was more specific. It therefore offers no general diagnostic advantage, although there may be some pigs in which infection can be detected by lgM EIA although they are MAT-negative.Furthermore, we have found IgM EIA to be of value for distinguishing antibodies produced in response to infection from maternally-derived antibodies, as the latter seem to be entirely IgG. White spots are well recognized as a sign of porcine leptospirosis at slaughter. Peet et al. (1983) isolated leptospires from 27 kidneys condemned at a Western Australian abattoir because of grossly visible lesions. Mercy et al. ( 1988 ) reported the culture of leptospires from 35 of 42 kidneys condemned in the same State. Hunter et al. (1987) in South Africa isolated serovar p o m o n a from 19/21 kidneys with white spots and from 1/9 apparently normal kidneys. They considered that leptospirosis was a more importance cause of macroscopic kidney lesions than a number of other bacterial pathogens. Baker et al. ( 1989 ) observed white spots in 63 (32%) of 197 randomly selected pigs slaughtered in southern Ontario, Canada, and showed associations with MAT titres to serovar p o m o n a and with the demonstration of renal leptospires by immunofluorescence. In an associated case-control study of 61 pigs, these authors demonstrated a relationship between white spots and the isolation of serovar p o m o n a . White spots have been reported to develop between one and three weeks post-infection (Burnstein and Baker, 1954; Sleight et al., 1960). We have observed experimentally infected pigs killed three weeks postinfection, with demonstrated kidney infection and with visibly normal kidneys (unpublished observations). We have also observed pigs (Jones et al., 1987) with visible lesions, in which infection has ceased. Clearly white spots may be
162
R.J. C H A P P E L ET AL.
misleading as a diagnostic indicator either early in infection or after infection has ceased. In addition, organisms other than leptospires may also sometimes cause similar lesions (Hunter et al., 1987). Vaccination of growing pigs has been reported to suppress white spots without always preventing renal infection (Cargill and Davos, 1981 ). Maternally-derived antibody may have a similar effect (Jones et al., 1987). In the present study we have confirmed the relationship between white spots and leptospiral infection using several diagnostic criteria (Table 4). However, many kidneys with white spots carried no evidence of infection. Large white spots were associated with leptospiral infection (1
151
Comparison of diagnostic procedures for porcine leptospirosis R.J. Chappel a, R.W. Prime b, B.D. Millarb, L.J. Mead a, R.T.
Jones b and
B. A d l e r c
aDepartment of Agriculture, Victorian Institute of A nimal Science, 475 Mickleham Road, A ttwood, Victoria 3049, Australia hDepartment of Agriculture, Regional Veterinary Laboratory, Bendigo, Victoria 3550, Australia CDepartment of Microbiology, Monash University, Clayton, Victoria 3168, Australia (Accepted 5 August 1991 )
ABSTRACT Chappel, R.J., Prime, R.W., Millar, B.D., Mead, L.J., Jones, R.T. and Adler, B., 1992, Comparison of diagnostic procedures for porcine leptospirosis. Vet. Microbiol. 30:151-163. Kidneys and matched serum samples were obtained from 368 pigs slaughtered at three Victorian abattoirs, and originating from 42 farms. Macroscopic lesions (white spots) were observed on 102 of the kidneys. Serum samples were tested by the microscopic agglutination test (MAT) and by an IgM enzyme immunoassay (EIA). Kidneys were cultured for leptospires, examined histologically after Warthin-Starry silver staining and after immunogold silver staining (IGSS), and tested for leptospiral DNA by DNA hybridization. Forty-four infected pigs were identified by culture or immunogold silver staining of kidneys or by high MAT titres (~> 1024). Infection was demonstrated in 7.5% of visibly normal kidneys, in 23.5% of kidneys with white spots, and in 48% of kidneys with large white spots, of I cm diameter or greater. The apparent (maximum) sensitivities of diagnostic procedures for detecting infection were as follows: MAT (at a titre of either 64 or 1024) 95%; IgM EIA 82%; culture 61%; presence of white spots 55%; IGSS 52%; presence of large white spots 30%; WarthinStarry silver staining 20%. IGSS, Warthin-Starry staining and DNA hybridization all appeared to be highly specific. Of 22 kidney sections identified as positive by IGSS, 13 showed intact leptospires, and these kidneys were all culture-positive. Nine others showed leptospiral antigen in the kidney tubules but no intact leptospires. Only five of these kidneys were culture-positive.
INTRODUCTION
Leptospirosis is a significant disease of pigs throughout the world. Leptospira interrogans serovar pomona is the most important serovar affecting pigs, and is responsible for abortions and stillbirths following the infection of susceptible gilts and sows. Such reproductive losses can be controlled by vaccinating breeding females with killed leptospiral vaccines. However leptospirosis can be maintained in growing pigs destined for slaughter, despite the vaccination of the sows from which they are derived. Infected grower pigs 0378-1135/92/$05.00 © 1992 Elsevier Science Publishers B.V. All rights reserved.
152
R.J. CHAPPEL ET AL.
present a h u m a n infection hazard at the abattoir, and meat inspectors and abatoir workers often contract leptospirosis from slaughtered pigs. Infected grower pigs are asymptomatic and cannot usually be identified when they are inspected before slaughter. Unless laboratory tests are performed, the only indication that a slaughtered pig may be infected with leptospires is the presence in some such animals of visible multifocal white lesions in the renal parenchyma ("white spots") (Jones et al., 1987; Hunter et al., 1987; Baker et al., 1989). These lesions are a reflection of interstitial nephritis, which is a c o m m o n result of kidney infection by leptospires. It is important to evaluate the specificity and sensitivity of laboratory diagnostic procedures. Furthermore, because white spots are commonly assumed to indicate leptospirosis, their value as a diagnostic indicator is also of interest. The purpose of this work was twofold. The primary aim was to obtain information on the diagnostic value and proper interpretation of several techniques used for the diagnosis of porcine leptospirosis. The second aim was to obtain information about the relationship between leptospiral infection and white spots observed at slaughter. MATERIALS AND METHODS
Collection and processing of kidneys and blood samples A single kidney with a paired blood sample was collected from each of 368 pigs at three Victorian abattoirs (at Altona, Castlemaine and Daylesford) between March 1988 and March 1989. Blood was collected at the point of slaughter, and the pig was marked with an individual tattoo which was subsequently used to identify the carcass when the kidney was collected at another part of the abattoir. The kidneys were not a random sample but were designed to provide a representative cross-section of kidneys with white spots and visibly normal kidneys. Collection was thus deliberately biased towards kidneys with white spots. No more than 10 kidneys with white spots and 10 visibly normal kidneys were accepted from any one farm.
Grading of kidneys for macroscopic lesions Kidneys were considered to be positive for white spots if they had at least two visible lesions. Kidneys with white spots were classified as having large lesions (about 1 cm diameter or greater) or having small lesions (generally pinhead size or smaller). Individual kidneys which had both large and small lesions were classified as having large lesions. All kidneys were graded by one veterinary pathologist.
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSP1ROSIS
15 3
Culture Kidneys were brought to the laboratory and cultures were set up as soon as practicable, always on the day of collection. A piece of each kidney about 1 cm 3, containing both cortex and medulla, was obtained aseptically from the inside of the kidney, and homogenized in a stomacher in 3 ml of sterile EMJH base. Three drops of the resulting homogenate were added to 5 ml of EMJH m e d i u m (Turner, 1970), supplemented with sodium pyruvate and sodium acetate (each 0.1 g/l) and with added 5-fluorouracil and actidione (each 0.1 g/l), and 3-drop serial transfers were then made into two further 5 ml volumes of medium. Cultures were incubated at 30 ° C, and the contents examined by dark field microscopy after 1-2 weeks, 6 weeks, 10 weeks and approximately 5 months. Kidneys were excluded from the study if all three cultures became contaminated by the 10 week inspection.
Histology Sections of the same piece of kidney, fixed in 10% buffered neutral formalin and processed in paraffin, were used for Warthin-Starry silver staining and for immunogold silver staining. Silver staining by the Warthin-Starry technique (Warthin and Starry, 1920) was performed as modified by Young ( 1969 ). One veterinary pathologist examined all slides under code. Immunogold silver staining (IGSS) of sections of formalin-fixed paraffinembedded kidney was performed as previously described (Skilbeck and Chappel, 1987). Slides were stained by one worker but were examined independently by two workers. Negative and positive control kidneys were included in each batch of slides processed.
DNA hybridization DNA hybridization ( D N A H ) , using a genomic probe derived from serovar pomona, was performed by a modification of the method of McCormick et al. (1989). Approximately 2 g of kidney tissue was homogenized in 2 ml of phosphate-buffered saline (PBS), pH 7.3 and samples were centrifuged at 800 g for 15 min. Equal volumes ( 500/4) of the clear supernatant and 0.8 M NaOH containing 20 m M EDTA were mixed and boiled for 10 min. To 500 fll of the supernatant after boiling, 50/zl of 3 M sodium acetate were added and the solution mixed. One volume of cold 95% ethanol was then added and the tube was shaken. Sample tubes were centrifuged for 5 min in an Eppendorf centrifuge, and the supernatant was removed twice using a vacuum pump. Samples were then resuspended in 10/A of TE buffer and heated at 60 ° C for 15 min, then a 3/tl sample was spotted onto a nylon filter had been prewetted with 6 X SSC and dried. The filter was then placed on blotting paper wet with 6 X SSC and exposed to u.v. light for 2 min. Hybridization was then performed as described by McCormick et al. (1989).
154
R.J. CHAPPEL ET AL.
Sixty-four of the kidneys were tested. Of these, 30 had infection proven by culture or by IGSS. Of the remainder, 22 were from herds from which the kidneys sampled had no proven infection and no white spots, and from which all sera were MAT-negative. The last 12 were from herds in which some of the pigs sampled had positive MAT titres a n d / o r white spots.
Microscopic agglutination test The microscopic agglutination test (MAT) (Faine, 1982 ) was performed using viable serovar pomona as antigen, on sera at final dilutions from 1/4 to 1/8192. Enzyme immunoassay An enzyme immunoassay for IgM antileptospiral antibodies (IgM EIA) was modified from that described by Ballard et al. (1984). Leptospires of serovar pomona at a concentration of 3 X 107/m| were disrupted by sonication and autoclaved to produce the EIA antigen. The wells of U-bottomed polystyrene plates (Nunc Polysorp ) were coated for 16-40 h at 4 oC with 100 ~tl volumes of EIA antigen. 100 ~1 of serum diluted in PBS-Tween containing 1 m g / m l bovine gamma globulin, and 100 #1 of peroxidase-conjugated antiporcine IgM (/~-chain specific) (Kirkegaard & Perry Laboratories) were diluted in PBS-Tween and were consecutively incubated at 37°C for 1 h. Substrate solution (100/zl) consisting of 1 m M 2,2'-azino-bis(3-ethylbenzthiazoline-6-sulphonic acid) (ABTS) in 0.10 M citrate-phosphate buffer pH 4.2, containing 0.009% hydrogen peroxide was added for 1 h at room temperature and the absorbance was read at 414 nm. Sera were tested in duplicate and were considered positive if the mean absorbances exceeded the mean absorbance of eight replicates of a standard serum containing 64 units of arbitrarily-defined antibody activity. A serum sample obtained from a young pig ten days after experimental infection with L. interrogans serovar pomona was defined as having 1024 units of activity. This sample was rich in IgM antibody, and was diluted in preinfection serum to obtain the 64 unit standard. The m i n i m u m diagnostic value of 64 units was adopted after testing sera from four experimentally-infected pigs (Ballard et al., 1984) using standard curves, and this value provided a compromise between false positive and false negative diagnostic reactions in those animals. All positive results were confirmed by retesting in another assay. Statistical analysis The analysis of ordinal categorical data was done by logistic regression using cumulative logits (McCullogh, 1980). Rating agreement between tests was measured by Cohen's Kappa test (Fleiss, 1973 ), in which K< 0 indicates less than chance agreement, K> 0 indicates more than chance agreement, and K= 1 indicates perfect agreement.
155
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSP1ROSIS
RESULTS
Kidneys and sera collected Kidneys and sera were obtained at three abattoirs from 42 farms, with from 1 to 20 kidneys coming from each individual farm (mean 8.8; median 10). A total of 102 kidneys with white spots was obtained, and these were derived from 28 of the farms. Of the 102 kidneys with white spots, 27 had large lesions ( 1 cm or greater), and these were derived from 13 farms.
Definition of an infected pig After analysis of the results presented below, individual pigs were considered to be infected if leptospires were demonstrated in the kidney by culture of IGSS, or if they had MAT titres I> 1024. A herd was considered to be infected if it contained at least one infected animal.
Culture A total of 27 out of the 368 kidneys was culture-positive (Table 1 ). Of these 27, leptospires were first seen after 1-2 weeks in 11, after 6 weeks in 15, and after 10 weeks in one. No additional positives were detected at the 5 months examination. There was no consistent pattern as to which dilution or dilutions became positive first, but in some cases all three dilutions had visible leptospires after 1-2 weeks. The culture-positive kidneys were derived from five farms of the 42. All were among the 28 farms from which kidneys with white spots were obtained, and four of the five were among the thirteen herds from which kidneys with large white spots were obtained. TABLE 1 R e s u l t s o b t a i n e d ~ o m 44 p i g s c l a s s i f i e d as p o s i t i v e b y c u l t u r e , i m m u n o g o l d s i l v e r s t a i n i n g , o r M A T 1024 Culture
IGSS
MAT
EIA
17
+
+
) 1024
+
12 5 2 2 2 1 1
+ + + -
+ + + -
) ) ) ) ) ) ~
+ + + -
1
-
+
1
+
-
No.
o f pigs
1024 1024 1024 1024 1024 1024 1024 32 0
-
156
R.J. CHAPPEL ET AL.
Immunogold silver staining A total of 21 kidneys was found to be positive by both workers, and one additional positive was detected by each, giving 23 IGSS-positive kidneys in all (Table 1 ). There was good agreement between the two workers (K= 0.84, P / = 1024. 2A herd was regarded as infected if it was shown to have one or more pigs regarded as infected. TABLE 3
Relationship between the results of the MAT and of the IgM EIA in individual sera IgM EIA
Negative Positive
Numbers of Sera MAT negative
MAT 4-256
MAT >/1024
195 73
39 19
6 36
White spots Table 4 compares the diagnostic results obtained with pigs with large white spots, with white spots of any size, or with visibly normal kidneys. Whereas white spots in general, and large white spots in particular, were clearly associated with leptospiral infection, white spots were not a particularly good predictor of current infection ( x = 0.18 ). Only 23.5% of pigs with kidneys with white spots were shown to be infected, and 7.5% of pigs with visibly normal kidneys were also shown to be infected. Large white spots were a better predictor of infection than white spots in general, and 48% of all pig kidneys with these spots were shown to be infected. Likewise 44% of pigs with large white spots on their kidneys had serum MAT titres of 1024 or above, a serological result shown above to be strongly associated with current infection. Among the 44 infected pigs, white spots were present on the kidneys of only 24 (55%), and large white spots on the kidneys of 13 (30%). A total of 14 of the 42 herds sampled in this study yielded no kidneys with white spots, and 116 pigs were sampled from these 14 herds. None of this group of pigs was
158
R.J. CHAPPEL ET AL.
TABLE 4 Diagnostic results obtained with pigs whose kidneys had large white spots, had white spots of any size. or were visibly normal Diagnostic result Total number Infected 3 Culture-positive IGSS positive total intact leptospires antigen only Warthin-Starry positive MAT>~4 M A T ) 1024 EIA positive
Large white spots 27 13 (48%) 8 (30%) 9 (33%) 3 (11%) 5 (19%) 3 (11%) 14 (52%) 12 (44%) 17 (63%)
Any white spots 102 24 (24%) 16 (16%) 16 (16%) 8 (8%) 7 (7%) 6 38 23 46
(6%) (37%) (23%) (45%)
No white spots 266 20 (8%) 11 (4%) 7 (3%) 5 (2%) 2 (1%) 3 62 19 82
(1%) (23%) (7%) (31%)
All pigs 368 44 (12%) 27 (7%) 23 (6%) 13 (4%) 9 (2%) 9 100 42 128
(2%) (27%) (11%) (35%)
3pigs w e r e considered to be infected if their kidneys were culture-positive or positive to IGSS, or if they had MAT titres > / = 1024.
shown to be infected. However 18 of them (16%) had serum antibody detectable by MAT (titres from 4 to 256), and 31 (27%) were positive by IgM EIA. DISCUSSION
Of the 368 pigs studied, 44 (12%) were considered to be infected at the time of slaughter. These included 27 whose kidneys were culture-positive, four whose kidneys were culture-negative but positive by IGSS, and a further 13 which had MAT titres >1 1024. It is possible that additional pigs were infected at the time of slaughter but were not readily detected by either culture or IGSS or by MAT, because other studies have shown the presence of infection in the kidneys of serologically negative pigs (Sturdza et al., 1969; Whyte and Ratcliff, 1982; Whyte et al., 1982). The studied population was drawn from 42 farms but was not randomly selected, as there was a deliberate attempt to collect kidneys with white spots. The proportion of kidneys with white spots that pass through Victorian abattoirs is not known, but is probably less than the 28% (102 out of 368) included in this study. Approximately 14% of pig kidneys inspected in South Australian abattoirs have white spots (P.R. Davies, personal communication). In this study, 24 of the 102 kidneys with white spots (23.5%) had proven infection, as against 20 of the 266 visibly normal kidneys (7.5%). The latter figure provides a lower limit on the prevalence of infection in the population studied. If 10% of all kidneys passing through the abattoirs studied are assumed to have white spots, the minimum level of infection can be esti-
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSPIROSIS
| 59
mated as 7.5 × 0.90 + 23.5 X 0.10 = 9%. Thus 9% or more of pigs slaughtered at these Victorian abattoirs during the study period were probably infected. A proportion of the animals studied had probably been infected during their lifetime, but kidney infection had ceased by the time of slaughter. The duration of shedding of leptospires in the urine of infected pigs is highly variable and may extend from a few days up to a year or more, but a few weeks may be most usual (Ryley and Simmons, 1954; Ryley, 1956; Mitchell et al., 1966; Edwards and Daines, 1979; Chappel et al., 1985 ). Thus a pig that is infected before three months of age may no longer be infectious at slaughter at about five months of age. It can nevertheless be expected to have a significant MAT titre at slaughter (Ballard et al., 1984), and some MAT-positive pigs in this study may therefore have been infected in the past. In this study 100 pigs had serum MAT titres from 4 to 8192 or above. If we discount the lowest group of titres, from 4 to 16, to eliminate any suggestion that the titres did not represent specific antibody attributable to infection, 65 titres of 32 and above remain to indicate that 18% of the pigs studied were infected or had been infected during their lifetime. Of these 65 pigs from 17 herds, 43 were infected, another three were pigs in proven infected herds, and 19 were from herds not known to be infected.
The value of different diagnostic procedures Evaluation of sensitivity requires the identification of a known positive group of animals, such as the 44 identified in this study (Table 1 ). Table 5 shows estimates of the apparent sensitivities (Martin, 1977) of the procedures used. In this study it was established that false negative diagnoses can be given by culture, IGSS, the Warthin-Starry technique, DNAH and IgM TABLE5 Apparent sensitivity of diagnostic procedures for detecting pigs proven to be infected in this study. These figures may be overestimates, as some infected pigs may not have been identified Diagnostic Procedure
Number of Infected Pigs Positive
Apparent Sensitivity
MAT (/>4) MAT (>/64) MAT ( >~1024) IgM EIA Culture White spots IGSS Large white spots Warthin-Starry DNAH
43 42 42 36 27 24 23 13 9 (7/30)
98% 95% 95% 82% 61% 55% 52% 30% 20% (23%)
160
R.J. CHAPPEL ET AL.
EIA. Other studies have shown that the MAT can also be negative in pigs infected with serovar p o m o n a (Sturdza et al., 1969; Whyte and Ratcliff, 1982; Whyte et al., 1982 ), and in this study one infected pig was serologically negative. It was therefore not possible to identify a group of pigs that were uninfected from the population studied, in order to derive numerical estimates of specificity. It was however apparent that the specificities of the Warthin-Starry technique, IGSS and DNAH were all excellent. Culturing detected 61% of known infected pigs over a period up to ten weeks. By definition it does not give false positive diagnostic results. IGSS detected 52% of known infected pigs. It appeared to give no false positive results. In 14 of 22 cases leptospires seen were visibly intact on IGSS. In each case the kidney was also culture-positive. In the other nine cases, of which only five were culture-positive, black dots were visible in the tubules, although absent from negative controls. Scanziani et al. (1989) suggested that such lesions represent later stages of infection, in which disintegration both of leptospires and of the tubules themselves can be observed. The IGSS was shown to be more sensitive than Warthin-Starry silver staining, and of high specificity. It has a role to play in the diagnostic laboratory as a more rapid alternative to culture. The well-established Warthin-Starry technique, although specific, was of low sensitivity in this study, detecting only 20% of infected animals. In our hands IGSS was a more satisfactory alternative. Hunter et al. (1987) failed to detect leptospires by the Warthin-Starry technique in any of 20 culturepositive kidneys. DNAH using a genomic probe derived from serovar pomona, originally applied successfully to urine (Millar et al., 1987), was of low sensitivity but specific when applied to kidney tissue. The sensitivity figure obtained is not directly comparable with the others, being drawn from a subgroup of the infected pigs. The results obtained are in agreement with those of McCormick et al. (1989), who found the technique to be of similar sensitivity to Warthin-Starry staining, but much less sensitive than culture. The culture-positive kidneys studied by McCormick et al. were among those collected in this study, but the technique used was not identical. DNAH has a number of general advantages in leptospirosis diagnosis. Like IGSS, it is much faster than culture, but unlike IGSS it can be readily applied to the simultaneous assay of large groups of samples. Although low sensitivity was observed in this study, adoption of the technique of polymerase chain reaction technology as reported for the diagnosis of bovine leptospirosis (Van Eys et al., 1989) may make DNAH the most sensitive technique available. The MAT was of high sensitivity. Although MAT-negative infected pigs have been reported in the literature, only one such animal was found in this study. Almost all infected pigs had high MAT titres. The significance of low MAT titres (256 or less) is not completely clear. Some of the pigs concerned may have been infected in the past, or may have been in early infection before
COMPARISON OF DIAGNOSTIC PROCEDURES FOR PORCINE LEPTOSPIROSIS
161
kidney infection was established. Many low titres (4 to 256) were observed in sera from herds not proven to be infected (Table 2 ), and it is possible that infection was present in these herds. When this work was started it was anticipated that, as antibodies detected by IgM EIA were shorter-lived in the sera of experimentally-infected pigs than were MAT titres (Ballard et al., 1984), the IgM EIA would give a more specific indication of current infection that the MAT. It was therefore expected that fewer sera would be positive by the IgM EIA than to the MAT, but this proved not to be the case. The work of Ballard et al. (1984) was performed in pigs experimentally-infected at eight weeks of age, which is several weeks earlier than pigs commonly become infected in the field in Victoria and South Australia (P.R. Davies and R.W. Prime, unpublished observations). From our unpublished field observations, infection at a greater age seems to produce a more prolonged IgM antibody response. The relationship between the results of the IgM EIA and those of the MAT (Table 3 ) was poor (~c= 0.26, P < 0.001 ). The MAT detects agglutinating surface epitopes, predominantly on the lipopolysaccharide, whereas EIA also detects antibodies against subsurface epitopes. The IgM EIA was less sensitive than the MAT in detecting infected pigs, and there was no evidence that it was more specific. It therefore offers no general diagnostic advantage, although there may be some pigs in which infection can be detected by lgM EIA although they are MAT-negative.Furthermore, we have found IgM EIA to be of value for distinguishing antibodies produced in response to infection from maternally-derived antibodies, as the latter seem to be entirely IgG. White spots are well recognized as a sign of porcine leptospirosis at slaughter. Peet et al. (1983) isolated leptospires from 27 kidneys condemned at a Western Australian abattoir because of grossly visible lesions. Mercy et al. ( 1988 ) reported the culture of leptospires from 35 of 42 kidneys condemned in the same State. Hunter et al. (1987) in South Africa isolated serovar p o m o n a from 19/21 kidneys with white spots and from 1/9 apparently normal kidneys. They considered that leptospirosis was a more importance cause of macroscopic kidney lesions than a number of other bacterial pathogens. Baker et al. ( 1989 ) observed white spots in 63 (32%) of 197 randomly selected pigs slaughtered in southern Ontario, Canada, and showed associations with MAT titres to serovar p o m o n a and with the demonstration of renal leptospires by immunofluorescence. In an associated case-control study of 61 pigs, these authors demonstrated a relationship between white spots and the isolation of serovar p o m o n a . White spots have been reported to develop between one and three weeks post-infection (Burnstein and Baker, 1954; Sleight et al., 1960). We have observed experimentally infected pigs killed three weeks postinfection, with demonstrated kidney infection and with visibly normal kidneys (unpublished observations). We have also observed pigs (Jones et al., 1987) with visible lesions, in which infection has ceased. Clearly white spots may be
162
R.J. C H A P P E L ET AL.
misleading as a diagnostic indicator either early in infection or after infection has ceased. In addition, organisms other than leptospires may also sometimes cause similar lesions (Hunter et al., 1987). Vaccination of growing pigs has been reported to suppress white spots without always preventing renal infection (Cargill and Davos, 1981 ). Maternally-derived antibody may have a similar effect (Jones et al., 1987). In the present study we have confirmed the relationship between white spots and leptospiral infection using several diagnostic criteria (Table 4). However, many kidneys with white spots carried no evidence of infection. Large white spots were associated with leptospiral infection (1
