J. Anim. Sci. 2003. 81:2401-2408
© 2003 American Society of Animal Science
Mortality of Creole kids during infection with gastrointestinal strongyles: A survival analysis1
N. Mandonnet*,2,
V. Ducrocq
,
R. Arquet
and
G. Aumont
* Institut National de la Recherche Agronomique, Station de Recherches Zootechniques, Domaine Duclos, 97170 Petit Bourg, Guadeloupe (French West Indies);
and
INRA, Station de Génétique Quantitative et Appliquée,78352 Jouy-en-Josas Cedex, France;
and
INRA, Domaine de Gardel, 97129 Moule, Guadeloupe(French West Indies);
and
INRA, Département de Santé Animale, 37380 Nouzilly, France
 |
Abstract
|
|---|
Mortality due to strongyles infection in small ruminants is a critical component of flock productivity in a tropical climate. In goat production, few experiments have been conducted to estimate the variability of this trait. A survival analysis study was carried out in the Creole experimental flock of INRA-Gardel (Moule, Guadeloupe) to identify management and genetic factors influencing mortality of kids reared at pasture and infected with gastrointestinal strongyles, predominantly Haemonchus contortus and Trichostrongylus colubriformis. Survival curves from 3 and 11 mo of age were analyzed for 837 kids sired by 48 bucks and 250 does. The causes of death were recorded. Mortality due to gastrointestinal strongyles was the variable considered. The flock management included drenchings with levamisole every 8 wk. Fecal egg counts and packed cell volume were regularly measured after 7 wk of natural infection. All but 6.7% of the records were uncensored, with an average failure time of 165 d. The probability of death following gastrointestinal infection was more than three times greater in males than in females. Kids raised by their mother before weaning had a lower (P < 0.05) relative risk of dying than those reared in nursery (0.40 vs. 1). Parity of the dam and litter size effects were not significant. The risk of death was reduced by approximately 80% during the 3 wk that followed a drenching (P < 0.01). Risk decreased by about 25% for each additional kilogram of body weight at weaning. Live weight, fecal egg counts, and packed cell volume all had significant effects on risk of death when introduced as time-dependent covariates in the model (P < 0.0001 for live weight and packed cell volume, and P < 0.01 for fecal egg counts). The estimated genetic variability was small and inaccurate. These results demonstrated that risk of death from gastrointestinal infection could be reduced with appropriate flock management. High infection levels increased the risk of death, but they were not the direct cause. The important mediation of reduced body weight and anemia in likelihood of death is highlighted. More data are needed to better assess the possibility for genetic improvement of viability in Creole kids during gastrointestinal strongyle infection.
Key Words: Genetic Variation • Goats • Mortality • Strongylidae • Survival
|
Introduction
|
|---|
Small ruminants are generally reported as successfully integrated with traditional production systems in the tropics (Ponzoni, 1988). However, mortality is a critical component of their productivity (Bradford, 1985
), and the occurrence of high mortality rates (e.g., due to disease) can dramatically reduce farmer incomes. Gastrointestinal nematode parasites remain one of the major constraints to small ruminant production in the tropics (Over et al., 1992). These infections are often identified as the main cause of mortality in sheep (Rege et al., 2002; Nguti et al., 2003) and in goat (Aumont et al., 1997b; Githigia et al., 2001).
This article reports a study of mortality in Creole goat, a highly prolific meat breed of the Caribbean characterized by Alexandre et al. (1999) and Mandonnet et al. (2002). Since genetic variability of resistance to gastrointestinal nematode parasites was assessed (Mandonnet et al., 2001), breeding for improved resistance in postweaning Creole kids is one way for reducing mortality. However, managing flock effects that increase risk of death could lead to more rapid improvement of tolerance in kids. Survival analysis methodology (Klein and Moeschberger, 1997) was considered to be the method of choice to assess variability of mortality in our experimental conditions. Since our main objective was to estimate genetic variability of resistance criteria, kids were drenched to reduce mortality rate. Even in situations of low incidence, such as ours, survival models proved to be efficient tools (Ducrocq et al., 2000). The rate at which the animals died is described without restricting the analysis to predefined cutpoints in time. The models use the available information, regardless of whether it comes from animals alive at the end of the period or from dead animals.
The aims of this study were to identify management factors influencing kids’ mortality due to infection with gastrointestinal strongyles and to estimate genetic parameters of this trait using survival analysis.
|
Materials and Methods
|
|---|
Experimental Design
The data of 837 kids born between 1995 and 1998, sired from 48 bucks and 250 does, were collected from the experimental Creole goat flock of INRA-Gardel, located in the French West Indies. An intensive reproduction system was applied (three parturitions every 2 yr). The kidding periods were the dry season (mid-February through mid-March), the intermediate season (mid-June through mid-July), and the wet season (mid-October through mid-November). The flock produced, on average, a 200-kid cohort every 4 mo. Throughout the year, the animals grazed on Digitaria decumbens irrigated pastures managed in a rotational system. Kids were weaned at an average age of 82 d. After weaning, males and females were grazed in separate paddocks for 8 mo (two seasons). The stocking rates ranged on average from 1.2 t of live weight/ha at the beginning of the period to 1.6 t/ha at the end. Sires’ pedigrees were traced back to the foundation in 1979. The causes of mortality were recorded. Death due to strongyles was the only source of mortality considered here.
Coccidiosis at weaning and infection by Moniezia during fattening were controlled by regular drenchings. Cowdriosis was completely controlled by twice-monthly accaricide application. Some cases of pneumonia and footrot were diagnosed during the wet season. Affected kids were removed from the experiment. Kids were naturally infected with gastrointestinal strongyles (mainly Haemonchus contortus and Trichostrongylus colubriformis). During the postweaning fattening period, kids were drenched every 8 wk with levamisole (12 mg/kg). Levamisole was effective (Barré et al., 1997) until the last three cohorts, when worm populations became resistant (G. Aumont, unpublished data).
Fecal samples were collected after 6 and 7 wk of each infection period (the interval between two drenchings) during fattening. Blood samples were collected for each animal every seventh week. Live weights were recorded at weaning, during fattening, at drenching, and in the middle of each infection period.
Fecal egg counts (FEC) were estimated using a modified McMaster method for rapid determination (Aumont et al., 1997a). In addition, fecal cultures were prepared to assess the composition of nematode burdens. Packed cell volume (PCV) was measured by the capillary microhematocrit method.
Survival Analysis
Theory.
For a complete description of theory, see Kalbfleisch and Prentice (1980) or Klein and Moeschberger (1997). The survival data of infected kids were described with a proportional hazards model, for which the hazard function h(t;xm) at time t of the animal m is written as follows:
 | [1] |
where t is time in days since weaning, h0(.) is the baseline hazard function and describes the overall risk of dying of the population, xm(t) is a vector of (possibly time-dependent) fixed covariates or indicator variables with ß as the vector of regression parameters, zm(t) is an incidence vector relating the hazard function to a vector of random (possibly correlated) effects. The function h0(.) can be left completely unspecified (defining the so-called Cox model) or may follow a parametric hazard, the most popular and flexible being the Weibull hazard function, (a two-parameter hazard distribution defined as h0(t)=
(t)-1). In order for the model to better fit the data, different baseline hazard functions h0,n(.) can be defined for each level (or stratum) n of a particular factor. The need for different baselines can be assessed using graphical tests. When the plots of the natural logarithm of [-ln
0,n(t)] against the natural logarithm of t (ln t) are parallel lines, a unique baseline can be assumed overall strata. Furthermore, a Weibull proportional hazard model can be assumed if a straight line is obtained.
Model selection.
The fixed effects of sex (male or female), parity of the dam (lactation one or two, lactation three to five, and lactation six and more), cohort (nine levels from 1995-1 to 1997-3), litter size (one, two, or three and more), rearing mode (maternal or artificial), and weight at weaning (as a continuous covariate) were tested using likelihood ratio tests. Only significant effects were included in the final model. The baseline hazard function was initially stratified per level of each effect, except the cohort effect and the effect of weight at weaning. Then the values of ln[-ln0,n(t)] were plotted against ln t to decide whether these baselines could be grouped together. Treatment application was modeled as a time-dependent variable (level = 1 during the 3 wk following each drenching, level = 2 after the 3 wk following each drenching). The effects of live weight, FEC, and PCV were also tested as time-dependent continuous covariates. Finally, sire effects were added and were assumed to follow a multivariate normal distribution. The sire variance was estimated using a Bayesian approach (Ducrocq and Casella, 1996). The characteristics (mean, mode, standard deviation) of the approximate marginal posterior density of this sire variance were calculated. All computations were done using Survival Kit version 3.12, a set of FORTRAN programs written with animal breeding applications in mind (Ducrocq and Solkner, 1998).
|
Results
|
|---|
From the 837 records analyzed, 93.3% were censored at an average time of 270 d (animals still alive at the end of the fattening period). Only 6.7% of the records were uncensored, with an average failure time of 165 d. Proportionally, the main strongyles species during the first 2 yr was H. contortus, but T. colubriformis predominated the last year. The infection level in the flocks ranged between 106 and 5,245 eggs per gram of feces. The PCV ranged from between 6 to 38% (24.7% on average).
Model Selection and Fixed Effects Estimates
Baseline survivor functions for the effect of rearing mode were grouped together because the plots of ln[-ln0,n(t)] against ln t appeared to be roughly parallel. Lines were not parallel for the sex effect, requiring the consideration of two different baselines (Figure 1). However, these lines are relatively straight, indicating that a Weibull model stratified by sex could be assumed. For each level of sex, the two baseline parameters and ln (= intercept) graphically estimated from the Cox model or directly estimated by maximum likelihood techniques for the Weibull model are presented in Table 1. They appear to differ when one or the other model is used, but this is partly the consequence of a strong negative correlation between the estimates of and ln . Because fewer a priori assumptions are postulated, only the estimates obtained from the Cox model are presented (Table 2) and discussed.
View larger version (17K):
[in this window]
[in a new window]
|
Figure 1. Graphical test of the assumption that the baseline hazard functions for infected males and females kids with strongyles during the fattening period are proportional and/or are Weibull hazards.
|
|
View this table:
[in this window]
[in a new window]
|
Table 1. Baseline Weibull parameters graphically estimated for Cox models and directly estimated for Weibull models, with or without pedigree information
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Estimates and hazard ratios (with 95% confidence intervals) for covariables affecting goat mortality due to endoparasites between 3 and 10 mo of age, estimated with the Cox model
|
|
Parity of the dam and litter size had no influence on probability of death following gastrointestinal infection during the postweaning period. Forty-eight of the 547 male kids died during the fattening period compared with eight of 290 in female kids. Two different patterns of baseline hazard rate were observed in the two sexes. This rate increased with time in males ( > 1) and decreased in females ( < 1). This corresponds to a death rate more than three times greater in males than in females (Figure 2). A trend was observed between seasons. The kids that were fattened during intermediate and wet season (1995-3, 1996-3, 1997-3) tended to have a higher probability of death than the others. However, an increased risk of death (P < 0.03) was observed in the last two cohorts (Table 2
) compared with the reference cohort (1995-1). Kids raised by their mothers before weaning had a lower (P < 0.05) hazard of mortality than kids artificially raised. The risk of death was reduced by almost 80% during the 3 wk that follow a drenching (P < 0.01). There was a linear relationship between weaning weight and risk of death (-0.029, P < 0.001). The estimated risk with the Cox model decreased by approximately 25% for each additional kilogram of BW at weaning. None of the interactions between two fixed effects was found to be significant.
View larger version (26K):
[in this window]
[in a new window]
|
Figure 2. Estimated survival curve of a) male and female kids; b) maternally and artificially reared kids; c) kids weaned with weights ranged from 6.5 to 12.5 kg; d) drenched or not drenched kids; e) kids bellowing to different cohorts.
|
|
When FEC was introduced as a time-dependent covariate in the model, the regression parameter was estimated as 0.146 (P < 0.01; not shown). This means that the risk of dying increased by 4.32 for every 1,000-unit increase in egg output. Live weight and PCV were also highly significant when included as time-dependent covariates in the model. The other effects of sex, rearing mode, weaning weight, drenching and cohort then became nonsignificant. The risk increased by 1.16 (-0.146, P < 10-4) and 1.69 (-0.523, P < 10-4) respectively for every unit decrease in PCV or kilogram decrease in live weight.
Estimated Survivor Curves
The estimated survivor curves are presented in Figure 2 for different populations characterized by different levels of sex, rearing mode, weaning weight, drenching, and cohort. Whatever the effect, survival rate mainly decreased at three steps: at the beginning of fattening (between 20 and 40 d), around 6 mo of age (between 100 and 120 d), and at the end of fattening (between 220 and 240 d). A proportion of 95% of females were still alive after 166 d vs. 119 d for males. The threshold of 90% of kids still alive was reached after 246 d and 115 d, respectively, for maternal and artificial rearing. After 260 d, the survival rate was about 15% higher for heavy kids (12.5 kg) at weaning than for light ones (6.5 kg). With drenching, 90% of the kids were still alive after 246 d. Without drenching, this mortality rate of 10% was reached 45 d earlier. In the different cohorts, the survival rates 260 d after weaning ranged from 97.1 to 85.5%.
Variance Components and Genetic Evaluation
Genetic parameters obtained with Cox models are reported in Table 3. As expected, the standard deviations and the skewness of the posterior density of sire effect were large and substantially changed when pedigree information was included. Using the formula of Yazdi et al. (2002), the heritability estimated with a sire model, in the unrealistic situation of no censoring, ranged from 0.50 to 0.70. After correction for the high censoring rate, the equivalent heritability (i.e., the heritability that can be used with selection index theory [Yazdi et al., 2002] to approximate reliabilities of genetic evaluations) was very low (around 0.05), but also very imprecise.
View this table:
[in this window]
[in a new window]
|
Table 3. Characteristics of the approximate marginal posterior density of the sire variance with a Cox mixed model with or without taking pedigree into account
|
|
|
Discussion
|
|---|
Herd management can be improved to reduce risk of death in kids during the critical period between weaning and 11 mo of age when the immune system is developing. The effect of sex on probability of death is consistent with the known higher resistance to strongyles in females (Barger, 1993). The risk of death was found to decrease in females and to increase in males during the fattening period. This is probably due to females more quickly acquiring immunity to strongyles than males between 3 and 11 mo of age. Special attention should be paid to the male flock in order to maintain lower infection rates. The adverse effect of the rainy seasons that we observed on survival rate of infected kids was previously reported by Nguti et al. (2003) in postweaning lambs under humid, tropical conditions. Alexandre et al. (1999) and Awemu et al. (1999) also reported increases in mortality rates of preweaning kids due to higher incidence of gastrointestinal strongyles during the wet season. Management practices should be adapted for this higher risk during the rainy period. The importance of drenching in terms of reducing mortality rate after infection is clearly illustrated and quantified. The limits of this control tool are also reached: The accumulation of worm burden at the end of our experiment (due to resistance to levamisole) at least doubled the risk of death in the two last cohorts. Controlling gastrointestinal strongyle infection in small ruminants only through anthelmintic application is dangerous and can lead to high mortality rates. Integrated control must be initiated, including, in particular, genetic improvement of resistance and tolerance traits.
The adverse effect of artificial rearing on survival is probably due to poorer grazing ability and fewer contacts with parasites before weaning. There is no evidence of protection against gastrointestinal nematode infections provided by the ingestion of colostrum from the dam (Dineen et al., 1978). Therefore, since parity of the dam and litter size were not significant, no maternal influence seems to modify survival of kids infected with gastrointestinal nematode parasites. These results are consistent with the lack of maternal effects reported for resistance criteria in Creole kids after 6 mo of age (Mandonnet et al., 2001). Postweaning survival and resistance can be evaluated on individual performances only. This suggests a good degree of adaptability to harsh tropical conditions in this local breed.
The inclusion of weight or PCV as a time-dependent covariate in the model eliminated the significance of the other effects. This result indicates that death in infected kids occurs after weakening (anemia) and severe loss of weight. High infection level (egg output) certainly increases the risk of death. However, it is not a direct cause of mortality. Some kids can harbor a high worm burden without dying. Very similar risk, of death per unit of PCV and FEC reported by Nguti et al. (2003). They also highlighted the important influence of reduced weight in likelihood of death. This result justifies the consideration of mortality as a resilience criterion compared with egg output, which is a resistance criterion.
Variability of small ruminant mortality is commonly analyzed as a binary trait (Lancelot et al., 2002; Rege et al., 2002; Baker et al., 2003). Currently no data has been published using survival analysis in kids. The major advantages of this methodology are the complete use of the information and the unique ability to incorporate covariates that vary with time, such as treatment, live weight, resistance criteria. Our experimental conditions were very similar to those found in clinical biometrics, where the methodology has primarily been developed, with a low rate of death and small data sets. Finally, it allowed for drawing conclusions on covariates of particular interest. A Cox model was used rather than a Weibull model because fewer assumptions are necessary.
The high censoring rate in our data set had two main consequences. First, it prevented us from correctly testing interactions between effects. Second, the level of mortality following strongyle infection was too low to express significant family differences. No genetic variability on mortality related to strongyles could be formally assessed from our data set on goat. Larger data sets are needed for a more precise assessment of genetic variability, but are not easy to obtain, at least when strongyle infection is to be studied in detail. However heritability estimates of mortality have generally ranged from 0 to 0.1 in sheep (Lopez-Villalobos and Garrick, 1999; Morris et al., 2000; Cloete et al., 2001), even in harsh environmental conditions (Burfening, 1993; Snyman et al., 1998). Therefore, little genetic progress should be expected for this critical component of flock productivity. When mortality is studied using survival analysis, higher genetic variability can be assessed (Nguti et al., 2003). Southey et al. (2001) reported estimates ranging from 0.1 to 0.2. They promoted selection to improve productivity and profitability of sheep production. Bishop et al. (2002) also suggested putting emphasis on this trait. Genetic variability in mortality would enable genetic improvement of tolerance to gastrointestinal strongyles infection and would therefore reduce the effects of disease in the flocks. Also, increased tolerance would not place much selective pressure on the pathogen, unlike resistance. At the moment, references on goat mortality are scarce and no genetic parameters have been previously published.
|
Implications
|
|---|
Overall, our results point out the main risk factors in goat kids infected with gastrointestinal nematode parasites. Appropriate flock management practices are suggested to decrease risk. Special care must be taken with male flocks and during the rainy season by reducing stocking rates on pasture or using mixed grazing systems. Drenchings need to be managed carefully through a monitor group or by selective treatment. Heavier live weights at weaning should be considered. Litter size must be appropriate for the dam’s milk production to decrease artificial rearing. These improvements in goat production are important factors for increasing productivity of small farms in the tropics. The survival analysis methodology was also successfully used to highlight the influence of reduced weight and weakening on the likelihood of death. However, further data are needed to draw conclusions on genetic improvement of viability in Creole kids during gastrointestinal strongyles infection.
|
Footnotes
|
|---|
1 This study was supported by the European Union (FEOGA) and "La Région Guadeloupe." 
2 Correspondence: INRA-URZ, Prise d’eau (phone: +33-590-25-54-08; fax: 33-590-25-59-36; E-mail: mandonne{at}antilles.inra.fr).
Received for publication October 17, 2002.
Accepted for publication June 3, 2003.
|
Literature Cited
|
|---|
Alexandre, G., G. Aumont, J. C. Mainaud, J. Fleury, and M. Naves. 1999. Productive performances of Guadeloupean Creole goats during the suckling period. Small Rum. Res. 34:155–160.
Aumont, G., R. Pouillot, and N. Mandonnet. 1997a. Le dénombrement des éléments parasitaires: Un outil pour l’étude de la résistance génétique aux endo-parasites chez les petits ruminants. Workshop final de l’ATP CIRAD-MIPA 72/94, Guadeloupe.
Aumont, G., R. Pouillot, R. Simon, G. Hostache, H. Varo, and N. Barré. 1997b. Intestinal parasitism of small ruminants in the French West Indies. Prod. Anim. (Paris). 10:79–89.
Awemu, E. M., L. N. Nwakalor, and B. Y. Abubakar. 1999. Environmental influences on preweaning mortality and reproductive performance of Red Sokoto does. Small Rum. Res. 34:160–165.
Baker, R. L., S. Nagda, S. L. Rodriguez-Zas, B. R. Southey, J. O. Audho, E. O. Aduda, and W. Thorpe. 2003. Resistance and resilience to gastrointestinal nematode parasites and productivity of Red Maasai, Dorper and Red Maasai x Dorper crossbred lambs in the sub-humid tropics. Anim. Sci. 76:119–136.
Barger, I. A. 1993. Influence of sex and reproductive status on susceptibility of ruminants to nematode parasitism. Int. J. Parasitol. 23:463–470.[Medline]
Barré, N., I. Amouroux, R. Aprelon, and T. Samut. 1997. Anthelmintic resistance of gastrointestinal nematodes in goat farms in Guadeloupe (French West Indies). Revue Elev. Méd. Vét. Pays trop. 50:105–110.
Bishop, S. C., J. Chesnais, and M. J. Stear. 2002. Breeding for disease resistance: Issues and opportunities. Communication No. 13-01 in Proc. 7th World Cong. Genet. Appl. to Livest. Prod., Montpellier, France.
Bradford, G. E. 1985. Selection for litter size. J. Anim. Sci. 49:66–74.
Burfening, P. J. 1993. Direct and maternal genetic effects on lamb survival. Small Rum. Res. 11:267–274.
Cloete, S. W. P., J. C. Greef, and R. P. Lewer. 2001. Environmental and genetic aspects of survival and early liveweight in Western Australian Merino Sheep. S. Afr. J. Anim. Sci. 31:123–130.
Dineen, J. K., P. Greeg, and A. K. K. Lascelles. 1978. The response of lambs to vaccination at weaning with irradiated Trichostrongylus colubriformis larvae segregating into ‘responders’ and ‘nonresponders.’ Int. J. Parasitol. 8:59–63.[Medline]
Ducrocq, V., B. Besbes, and M. Protais. 2000. Genetic improvement of laying hens viability using survival analysis. Genet. Sel. Evol. 32:23–40.[Medline]
Ducrocq, V., and G. Casella. 1996. A Bayesian analysis of mixed survival models. Genet. Sel. Evol. 28:505–529.
Ducrocq, V., and J. Sölkner. 1998. "The Survival Kit—V3.0," a package for large analyses of survival data, Pages 447–448 in Proc. 6th World Cong. Genet. Appl. to Livest. Prod., Armidale, Australia.
Githigia, S. M., S. M. Thamsborg, W. K. Munyua, and N. Maingi. 2001. Impact of gastrointestinal helminths on production in goats in Kenya. Small Rum. Res. 42:21–29.
Kalbfleisch, J. D., and R. L. Prentice. 1980. The Statistical Analysis of Failure Time Data. John Wiley and Sons. New York.
Klein, J., and M. Moeschberger. 1997. Survival Analysis. John Wiley and Sons, New York.
Lancelot, R., M. Lesnoff, and J. J. McDermott. 2002. Use of Akaike information criteria for model selection and inference: an application to assess prevention of gastrointestinal parasitism and respiratory mortality of Guinean goats in Kolda, Senegal. Prev. Vet. Med. 55:217–240.[Medline]
Lopez-Villalobos, N., and D. J. Garrick. 1999. Genetic parameter estimates for lamb survival in Romney sheep. Proc. N. Z. Soc. Anim. Prod. 59:121–124.
Mandonnet, N., G. Aumont, J. Fleury, R. Arquet, H. Varo, L. Gruner, J. Bouix, and J. Vu. Tien. 2001. Assessment of genetic variability of resistance to gastrointestinal nematode parasites in Creole goats in the humid tropics. J. Anim. Sci. 79:1706–1712.[Abstract/Free Full Text]
Mandonnet, N., A. Menedez-Buxadera, R. Arquet, M. Naves, and G. Alexandre. 2002. Genetic variability on post-weaning growth traits in Creole goats reared on pasture. Communication No. 11-26 in Proc. 7th World Cong. Genet. Appl. to Livest. Prod., Montpellier, France.
Morris, C. A., S. M. Hickey, and J. N. Clarke. 2000. Genetic and environmental factors affecting lamb survival at birth and through to weaning. N. Z. J. Agric. Res. 43:515–524.
Nguti, R., P. Janssen, G. J. Rowlands, J. O. Audho, and R. L. Baker. 2003. Survival of Red Maasai, Dorper and crossbred lambs in the humid tropics. Anim. Sci. 76:3–17.
Over, H.J., J. Jansen, and P. W. Von Olm. 1992. Distribution and impact of helminth diseases of livestock in developing countries. Paper 96 in Food and Agriculture Organisation of the United Nations Animal Production and Health.
Ponzoni, R.W. 1992. Genetic improvement of hair sheep in the tropics. Paper 101 in Food and Agriculture Organisation of the United Nations Animal Production and Health.
Rege, J. E. O., S. Tembely, E. Mukasa-Mugewa, S. Sovani, D. Anindo, A. Lahlou-Kassi, S. Nagda, and R. L. Baker. 2002. Effect of breed and season on production and response to infections with gastrointestinal nematode parasites in sheep in the highlands of ethiopia. Livest. Prod. Sci. 78:159–174.
Snyman, M. A., G. J. Erasmus, and J. B. Wyk. 1998. The possible genetic improvement of reproduction and survival rate in Afrino sheep using a threshold model. S. Afr. J. Anim. Sci. 28:120–124.
Southey, B. R., S. L. Rodriguez-Zas, and K. A. Leymaster. 2001. Survival analysis of lamb mortality in a terminal sire composite population. J. Anim. Sci. 79:2298–2306.[Abstract/Free Full Text]
Yazdi, M. H., P. M. Visscher, V. Ducrocq, and R. Thompson. 2002. Heritability, reliability of genetic evaluations and response to selection in proportional hazard models. J. Dairy Sci. 85:1563–1577.[Abstract]
Top
Abstract
Introduction
