A review of methods for estimating mortality due to parasites in wild fish populations

Distribution of different species of metacercariae in two freshwater fishes: Haludaria fasciata (Teleostei: Cyprinidae) and Pseudosphromenus cupanus (Teleostei: Osphromenidae)

Information on the distribution and abundance patterns of trematodes are essential to reveal the ecology of host-parasite interactions. The Western Ghats of India, a biodiversity hotspot, is rich in freshwater fish diversity and endemism. Though there are several studies on various other aspects of fish ecology, studies on their parasitic fauna is meager. The objective of the present study is to explore the distribution and infection patterns of metacercariae of five species of trematodes in the freshwater fishes, Haludaria fasciata and Pseudosphromenus cupanus. The infection parameters were analyzed for each host and CART model was applied to analyze the environmental factors affecting parasite distribution patterns. All species of metacercariae showed an over-dispersed aggregate distributions. The classification tree models indicated that among the environmental factors considered, differences in host locality was the most influential factor in both fishes, followed at a greater distance by the factor seasonality. The parasite communities exhibited temporal and spatial differences in the infection pattern in response to seasonal and locational variations.

Why ignoring parasites in fish ecology is a mistake

Parasites are ubiquitous components of biological systems that have evolved in multiple independent lineages during the history of life, resulting in a diversity of taxa greater than that of their free-living counterparts. Extant host-parasite associations are the result of tight reciprocal adaptations that allow parasites to exploit specific biological features of their hosts to ensure their transmission, survival, and maintenance of viable populations. As a result, parasites may affect host physiology, morphology, reproduction or behaviour, and they are increasingly recognized as having significant impacts on host individuals, populations, communities and even ecosystems. Although this is usually acknowledged by parasite ecologists, fish ecologists often ignore parasitism in their studies, often acting as though their systems are free of parasites. However, the effects of parasites on their hosts can alter variables routinely used in fish ecology, ranging from the level of individual fish (e.g. condition factors) to populations (e.g. estimates of mortality and reproductive success) or communities (e.g. measures of interspecific competition or the structure and functioning of food webs). By affecting fish physiology, parasites can also interfere with measurements of trophic levels by means of stable isotope composition, or have antagonistic or synergistic effects with host parameters normally used as indicators of different sources of pollution. Changes in host behaviour induced by parasites can also modify host distribution patterns, habitat selection, diet composition, sexual behaviour, etc., with implications for the ecology of fish and of their predators and prey. In this review, we summarise and illustrate the likely biases and erroneous conclusions that one may expect from studies of fish ecology that ignore parasites, from the individual to the community level. Given the impact of parasites across all levels of biological organisation, we show that their omission from the design and analyses of ecological studies poses real risks of flawed interpretations for those patterns and processes that ecologists seek to uncover.

Odds ratios and hurdle models: a long-term analysis of parasite infection patterns in endangered young-of-the-year suckers from Upper Klamath Lake, Oregon, USA

We used odds ratios and a hurdle model to analyze parasite co-infections over 25 years on >20,000 young-of-the year of endangered Shortnose and Lost River Suckers. Host ecologies differed as did parasite infections. Shortnose Suckers were more likely to be caught inshore and 3-5 times more likely to have Bolbophorus spp. and Contracaecum sp. infections, and Lost River Suckers were more likely to be caught offshore and approximately three times more likely to have Lernaea cyprinacea infections. An observed peak shift seems likely to be due to a lower host size limit for Bolbophorus spp. (13.6 mm) compared with L. cyprinacea (23.4 mm). The large data set allowed us to generate strong hypotheses: (i) that a major marsh restoration project had unintended consequences that resulted in an increase in infections; (ii) that co-infection with Bolbophorus spp. increased the odds of infection by L. cyprinacea and Contracaecum sp.; (iii) that significant declines in the odds of infection over approximately 25 days were due to parasite-induced host mortality; (iv) that the fish's small size relative to L. cyprinacea and Contracaecum sp. might be directly lethal; (v) that the absence of L. cyprinacea infections in the early 1990s was associated with good year-class production of the suckers; and (vi) that parasites might increase the odds of vagrancy from the nursery ground.

Distribution ofPomphorhynchus laevis s.l.(Acanthocephala) among fish species at a local scale: importance of fish biomass density

Parasite distribution among hosts is a fundamental aspect of host–parasite interactions. Aggregated parasite distributions within and across host species are commonly reported and potentially influenced by many factors, whether host or parasite specific, or related to host–parasite encounter and compatibility. Yet, the respective role of each in observed parasite distributions are often unclear. Here, we documented the distribution of the acanthocephalan parasitePomphorhynchus laevis sensu lato(s.l.) in two replicate fish host populations. Aggregated distributions were observed in both populations, within and across fish host species. We found a positive abundance–prevalence relationship across fish species, suggesting that resource availability (fish host biomass density) was the main driver ofP. laevis s.l.distribution. This was supported by further positive associations between mean parasite load and fish biomass density. We found little evidence for intensity-dependent regulation within host (i.e. intra-host competition among co-infecting parasites). Furthermore,P. laevis s.l.infection had no detectable effect on fish condition indices, except on the body condition of female barbel (Barbus barbus). Therefore,P. laevis s.l.tended to accumulate with size/age within fish species, and with fish biomass density among fish species, with apparently negligible limitations due to intra-host intensity-dependent regulation of parasite, or to parasite-induced morbidity in fish. The relative availability of final hosts for trophic transmission thus appears to be the main driver ofP. laevis s.l.distribution among fish.

The parasitic isopod Mothocya nana drives dietary shifts and poorer condition of Brazilian silversides Atherinella brasiliensis

Cymothoids are ectoparasites that may attach to various parts of the fish for molting and reproductive purposes, thus increasing the energetic costs of the host. This study investigated the influence of the parasitic isopod Mothocya nana on the physiological condition and diet of adult Brazilian silversides Atherinella brasiliensis at a sandy beach in southeastern Brazil. We collected 268 A. brasiliensis individuals, of which 230 fish were non-parasitized (mean ± SE total weight [TW] = 16.92 ± 0.38 g; total length [TL] = 127 ± 0.88 mm) and 38 were parasitized by up to 2 isopods (TW = 15.89 ± 0.79 g; TL = 126 ± 1.96 mm). Parasitic prevalence (P) and intensity (I) reached highest values in June 2015 (P = 20.88%; I = 1.31) and were slightly higher on males (P = 17.39%; I = 1.33) than on females (P = 13.07%; I = 1.27). Parasitized fish revealed poorer condition than non-parasitized ones, among which male hosts were especially burdened by M. nana's attachment. The condition factor, the eviscerated condition factor and the fullness index each showed a decreasing trend according to the parasite's development and offspring weight (i.e. increase in egg and larval weight). Parasitized and non-parasitized Brazilian silversides fed mainly on microcrustaceans, but the first group showed reduced phytoplankton intake and was associated with fewer trophic categories in comparison to non-parasitized fish. These dietary shifts revealed correlation with the poorer physiological condition reported for infected A. brasiliensis, whose decreased feeding efficiency is likely related to potential impairment of the filter-feeding mechanism and/or altered behaviour due to pressure atrophy and the increased energetic costs imposed by M. nana's development.

Histopathology and external examination of heavily parasitized Lost River Sucker Deltistes luxatus (Cope 1879) and Shortnose Sucker Chasmistes brevirostris (Cope 1879) from Upper Klamath Lake, Oregon

Shortnose Sucker (Chasimistes brevirostris) and Lost River Sucker (Deltistes luxatus) are endemic to the Upper Klamath Basin of Southern Oregon and Northern California, and their populations are in decline. We used histopathology and external examination of 140 and external examination only of 310 underyearling suckers collected in 2013, 2015 and 2016 to document pathological changes, particularly those relating to parasites. The most severe infection was caused by a Contracaecum sp., infecting the atrium of 8%-33% of Shortnose Suckers. The most prevalent infections were caused by Bolbophorus sp. metacercariae in the muscle of Shortnose Suckers (21%-63%) and Lernaea cyprinacea in the skin and muscle of Lost River Suckers (30%-81%). Histology detected Bolbophorus in only 5% of cases where it was not seen externally. Three myxozoans were observed; a Parvicapsula sp. in the renal tubules (10%), a Myxobolus sp. in the intestinal mucosa (2%) and an unusual multicellular, presporogonic myxozoan in the intestinal lumen of one sucker. Severe gill epithelial hyperplasia was observed in several fish collected in 2016. Trichodinids and Ichthyobodo sp. were observed on some of the gills, but absent in many of the fish with severe lesions. A histiocytic sarcoma was observed in sucker.

Hematology and condition factor of tui chub and fathead minnow parasitized by nematode from Upper Klamath Lake, Oregon, USA

This study evaluated the hematological profile and condition factor (Kn) of tui chub Siphateles bicolor and fathead minnow Pimephales promelas and their associations with larvae of Contracaecum sp. infection of the heart. A total of 30 tui chub and 17 fathead minnow were collected from Upper Klamath Lake, Oregon, USA, measured, and weighed and blood was drawn for hematological analysis. Nematode larvae parasitized tui chub with a prevalence of 50% and mean intensity of 1.40, while 11.8% of fathead minnow were parasitized at a mean intensity of 1.0. Non-parasitized tui chub were significantly larger than the parasitized fish, indicating that small fish could be easily predated by the definitive host, a piscivorous bird. Although the relatively large worm occupied a large portion of the atrium, the presence of the larvae did not affect tui chub Kn, possibly associated with low parasite intensity and a harmonic co-evolution. Only parasitized fathead minnow showed significant differences in red blood cell measurements (greater cell width and larger nuclei) compared to non-parasitized fish. Lymphocytes were the most common white blood cells found in tui chub, followed by neutrophils, monocytes, and periodic acid-Schiff positive granular leukocytes; in fathead minnow lymphocytes were followed by heterophils, monocytes, neutrophils and eosinophils. This study is the first report of Kn and description of blood cells and hematological parameters in these fish species.

Is an Apicomplexan Parasite Responsible for the Collapse of the Iceland Scallop (Chlamys islandica) Stock?

Due to the total and unexpected collapse of the Iceland scallop, Chlamys islandica, stocks around Iceland during the 2000s, a commercial fishing ban has been imposed on this valuable resource since 2003. Following the initial identification of an apicomplexan parasite in the scallops, a long-term surveillance program was established to evaluate the effect of the parasite on the population. The infections were highly prevalent in all shell sizes throughout the study. However, the parasite only impacts mature scallops where they cause severe macroscopic changes, characterized by an extensively diminished and abnormally coloured adductor muscle. A highly significant relationship was observed between infection intensity and gonad and adductor muscle indices. The first four years of the study, were characterized by high infection intensity and very poor condition of the adductor muscle and gonads, whilst during subsequent years, infections gradually decreased and the condition of the scallops improved. Histopathological changes were restricted to the presence of apicomplexan zoites which were widely distributed, causing varying degrees of pathology in all organs. In heavy infections, muscular and connective tissues were totally necrotized, destroying significant parts of numerous organs, especially the adductor muscle, digestive gland and gonads. The progression of the disease was in good synchrony with the mortality rates and the subsequent decline observed in the scallop stock and recruitment indices. Our findings strongly suggest that the apicomplexan parasite played a major role in the collapse of the Iceland scallop stock in Breidafjordur. In addition to causing mortality, the infections significantly impact gonad development which contributes further to the collapse of the stock in the form of lower larval recruitment. Furthermore, compelling evidence exists that this apicomplexan pathogen is causing serious disease outbreaks in other scallop populations. Similar abnormal adductor muscles and the parasite itself have been identified or observed in association with other mass mortality events in several different scallop species and commercial stocks in the northern hemisphere.

Detecting and quantifying parasite-induced host mortality from intensity data: method comparisons and limitations

Parasites can significantly impact animal populations by changing host behaviour, reproduction and survival. Detecting and quantifying these impacts is critical for understanding disease dynamics and managing wild animal populations. However, for wild hosts infected with macroparasites, it is notoriously difficult to quantify the fatal parasite load and number of animals that have died due to disease. When ethical or logistical constraints prohibit experimental determination of these values, examination of parasite intensity and distribution data may offer an alternative solution. In this study we introduce a novel method for using intensity data to detect and quantify parasite-induced mortality in wildlife populations. We use simulations to show that this method is more reliable than previously proposed methods while providing quantitative estimates of parasite-induced mortality from empirical data that are consistent with previously published qualitative estimates. However this method, and all techniques that estimate parasite-induced mortality from intensity data alone, have several important assumptions that must be scrutinised before applying those to real-world data. Given that these assumptions are met, our method is a new exploratory tool that can help inform more rigorous studies of parasite-induced host mortality.

Bacciger bacciger (Trematoda: Fellodistomidae) infection effects on wedge clam Donax trunculus condition

Wedge clams Donax trunculus inhabit high-energy environments along sandy coasts of the northeastern Atlantic Ocean and the Mediterranean Sea. Two sites were sampled monthly, one in Morocco (Mehdia), where the density was normal, and one in France (Biscarosse), where the density was very low. We tested the hypothesis that the difference in density between the sites was related to infection by the trematode parasite Bacciger bacciger. Identity of both the parasite and the host were verified using anatomical and molecular criteria. Parasite prevalence (i.e. the percentage of parasitized clams) was almost 3 times higher at Biscarosse. At this site, overall prevalence reached 32% in July and was correlated with the migration of several individuals (with a prevalence of 88%) to the sediment surface. After this peak, prevalence decreased rapidly, suggesting death of parasitized clams. The deleterious effect of B. bacciger on wedge clams was also supported by our calculations indicating that the weight of the parasite made up to 56% of the total weight of the parasitized clams. However, condition indices of trematode-free clams were also lower in Biscarosse than in Mehdia or other sites, suggesting that other factors such as pollutants or microparasites (Microcytos sp.) may alter wedge clam population fitness in Biscarosse.

Overwinter body condition, mortality and parasite infection in two size classes of 0+ year juvenile European bitterling Rhodeus amarus

Body condition and parasite abundance were examined in two size classes of European bitterling Rhodeus amarus during the first overwintering period in two seasons (2007-2008 and 2009-2010). Body condition of large fish did not change during winter, and increased significantly in March. From November to February, small fish showed a decreasing trend in condition. Despite a significant increase in March condition of small fish only reached the same level as before winter. Total parasite abundance increased significantly in winter in both fish size classes, reflecting a seasonal increase in monogenean infection. Large fish were parasitized significantly more than small fish during winter, but only in small fish was a negative correlation between parasite infection and condition found and a significant decrease in parasite abundance recorded after wintering, indicating mortality of heavily infected individuals with low condition during the winter. A trend for higher overwinter mortality in small fish was found under semi-experimental conditions. The decrease in condition during the winter period in small fish may reflect faster energy depletion generally expected in smaller individuals. The results indicate that parasite infection may contribute to the overwinter mortality of 0+ year R. amarus, with a stronger effect in smaller individuals.

Generation time and the stability of sex-determining alleles in oyster populations as deduced using a gene-based population dynamics model

Crassostrea oysters are protandrous hermaphrodites. Sex is thought to be determined by a single gene with a dominant male allele M and a recessive protandrous allele F, such that FF animals are protandrous and MF animals are permanent males. We investigate the possibility that a reduction in generation time, brought about for example by disease, might jeopardize retention of the M allele. Simulations show that MF males have a significantly lessened lifetime fecundity when generation time declines. The allele frequency of the M allele declines and eventually the M allele is lost. The probability of loss is modulated by population abundance. As abundance increases, the probability of M allele loss declines. Simulations suggest that stabilization of the female-to-male ratio when generation time is long is the dominant function of the M allele. As generation time shortens, the raison d'être for the M allele also fades as mortality usurps the stabilizing role. Disease and exploitation have shortened oyster generation time: one consequence may be to jeopardize retention of the M allele. Two alternative genetic bases for protandry also provide stable sex ratios when generation time is long; an F-dominant protandric allele and protandry restricted to the MF heterozygote. In both cases, simulations show that FF individuals become rare in the population at high abundance and/or long generation time. Protandry restricted to the MF heterozygote maintains sex ratio stability over a wider range of generation times and abundances than the alternatives, suggesting that sex determination based on a male-dominant allele (MM/MF) may not be the optimal solution to the genetic basis for protandry in Crassostrea.

A method for estimating possible parasite-related host mortality, illustrated using data from Callitetrarhynchus gracilis (Cestoda: Trypanorhyncha) in lizardfish (Saurida spp.)

The frequency distribution of parasites in hosts commonly follows a negative binomial or similar distribution. Under certain conditions the magnitude of parasite-associated host mortality can be estimated by comparing the tail of the observed distribution to that of the distribution predicted from the first few points of the data. For the technique to work the following assumptions need to be met: mortality in lightly infected fish must be rare; infection and consequent mortality occur only in fish younger than those sampled; and the frequency distribution of the parasite at the time of infection should conform to a known probability distribution. The method was applied to frequency distributions of blastocysts of Callitetra rhynchus gracilis in 898 Saurida tumbil (Bloch) and 5013 S. undosquamis (Richardson). Parasite-associated mortality in S. tumbil was calculated to be at least 11 % in males and 2% in females. For S. undosquamis, estimated mortality was about 5% in males and 3% in females. The numbers of parasites estimated to produce a 0·5 probability of death, the parasitological equivalent of an LD50 were 3·4 and 5·7 for S. tumbil males and females, and 18 and 3 for S. undosquamis males and females respectively.

Host preference of adult Caligus elongatus Nordmann in the laboratory and its implications for Atlantic cod aquaculture

The sea louse Caligus elongatus utilizes a range of marine fish species as hosts, and it has been suggested that fish farms are infected with adult C. elongatus from wild fish hosts. We assessed experimentally whether dislodged lice return to their original host species. Adult lice from wild lumpfish and saithe were presented to lumpfish, sea trout, Atlantic cod, saithe and European plaice. All fish species became infected. Lumpfish and cod were strongly preferred, followed by saithe, trout and plaice. Lice from wild saithe showed a preference for saithe and lumpfish, followed by cod, sea trout and plaice. The mtCO1 gene was used to determine the genotypes of the lice. Lice from lumpfish were genotype 1. Lice from saithe were mainly genotype 2. The two genotypes of C. elongatus differed slightly in their host preferences: lice from saithe were less discriminating. The data support the hypothesis that adult C. elongatus may transfer between different host species. Cod is an attractive host for lice from both lumpfish and saithe, which suggests that C. elongatus may become a serious pest of farmed cod.

Parasites of recruiting coral reef fish larvae in New Caledonia

Recruiting coral reef fish larvae from 38 species and 19 families from New Caledonia were examined for parasites. We found 13 parasite species (Platyhelminthes: Monogenea, Cestoda and Trematoda) but no acanthocephalan, crustacean or nematode parasites. Over 23% of individual fish were infected. Didymozoid metacercariae were the most abundant parasites. We conclude that most of the parasites are pelagic species that become 'lost' once the fish larvae have recruited to the reef. Larval coral reef fish probably contribute little to the dispersal of the parasites of the adult fish so that parasite dispersal is more difficult than that of the fish themselves.

The relationship between Pseudodiplorchis americanus (Monogenea) density and host resources under controlled environmental conditions

A previous study has shown that, under natural conditions, energy reserves of the desert toad, Scaphiopus couchii, are negatively related to the density of infection by Pseudodiplorchis americanus. However, this was based predominantly on collections of active animals from breeding congregations and inevitably selected toads which were in good condition. The parasite, a blood-feeding monogenean, occurs in burdens of up to 30 worms/host (mean intensity 6 worms/host) and represents a significant drain on reserves because the host does not feed during a 10-month hibernation. Field studies cannot resolve the possibility that larger worm densities are not observed in nature due to parasite-induced host mortality. The present study was conducted during investigations of P. americanus development and survival under controlled laboratory conditions, utilizing experimentally infected hosts which created worm densities larger than those observed in natural populations. At all temperature regimes, infected animals had smaller fat bodies than those uninfected but differences were generally not statistically significant due to large individual variations, presumably resulting from variations in past feeding efficiency. At cool temperatures (15-20 degrees C) there was no density-dependent effect on host fat body weight, and at a diurnal temperature cycle of 20-34 degrees C (simulating that experienced by host and parasite during the summer months), the effects of high temperatures were greater than the effects of infection, due to increased toad metabolic rates. The most significant effects of P. americanus were observed in hosts that began hibernation in relatively poor condition and experienced moderate temperatures (25 degrees C) during hibernation. The toads generally maintained packed blood cell levels (PCV) levels even when fat body weights were low, but infected animals had a lower PCV irrespective of fab body levels. In animals unfed after field collection, PCV was reduced in uninfected toads and was even lower in infected animals. Although very heavily infected toads (burdens of 35-95 worms/host) were generally in poorer condition than uninfected toads they still survived long-term hibernation under extreme nutritional stress. This study therefore confirmed observations made in field studies that there is a density-dependent relationship between the hosts' survival prospects and P. americanus infection. However, given the large variability in feeding efficiency and stored resources between individual toads, there is no evidence that the most heavily infected toads would have been unrepresented in field samples due to parasite-induced mortality.

Marine parasites: an Australian perspective

A review is given of major studies in marine parasitology in Australia. Aspects discussed include: geographical distribution of parasites in Australian coastal waters and their affinities to parasites of other zoogeographical regions; species diversity in Australian coastal surface and deep waters; use of marine parasites for stock discrimination; use of marine parasites as ecological models; ultrastructural and phylogenetic studies of marine parasites; and effects of marine parasites on their hosts.

The epidemiology of the metacercariae of Diplostomum baeri and D. spathaceum in perch (Perca fluviatilis) from the warm water effluent of a nuclear power station

Thermal effects on the dynamics of infection with metacercariae of Diplostomum baeri and D. spathaceum were monitored between May and September, in a year-class of perch Perca fluviatilis, in an artificial lake receiving warm water discharges from a nuclear power station and in an unheated reference site, for a two year period. In the heated area the prevalence of infection of the retinal form, D. baeri, was always 100%, whereas in the unheated site there was an increase from 93% in May to 100% in June 1986. The relative density of D. baeri was found to increase gradually during the first summer in both areas, although the accumulation rate of metacercariae was significantly increased in the heated area. In August of the first year the relative density of D. baeri peaked in the heated area, whereas it continued to increase in the unheated control. However, in September of the following year, the relative densities were at the same level in both thermal regimes. A concomitant decline in the degree of overdispersion of metacercariae within the host population was observed in the heated area as the population density of metacercariae decreased, whereas the index of dispersion remained at the same level throughout the study in the unheated area. However, it was not possible to sample perch in the unheated area between October 1986 and May 1987 and changes in the parasite population could have occurred during this period. The prevalence and relative density of D. spathaceum, the lens form, was on the contrary low, especially in the unheated site where it was recorded only occasionally. The infection of D. baeri exhibited a convex pattern in both thermal regimens, although the peak infection was noticed earlier in the heated area. Regulation of the parasite infrapopulation may have been achieved by the combined effects of a decreased transmission rate of cercariae with increasing age of the host, the mortality of metacercariae as a natural termination of the life span and to selective predation of heavily infected hosts. However, regardless of an increased accumulation rate of metacercariae in the heated area, the relative densities of D. baeri became equal in both thermal regimes at the end of the study. The mortality rate of metacercariae in the heated area was therefore presumably increased as compared with the unheated reference site.

The relation between the number of parasites/host and host age: population dynamic causes and maximum likelihood estimation

We examined dynamical factors that shape the distribution of the number of parasites/host in constant or temporally varying environments, and with or without host-age dependent variation in host susceptibility and parasite mortality. We predict properties of the parasite distribution in the absence of density-dependent factors such as density-dependent mortality of recruitment and parasite-induced host mortality. These properties provide a criterion for the detection of density dependence in temporally variable systems with host-age dependent interactions. We have then introduced methods to estimate and statistically evaluate the effects of host age or size on the distribution of parasites/host. The methods are based on a maximum likelihood protocol for linear and non-linear regression when data are negatively binomially distributed. We have illustrated the use of the theoretical results and statistical methods by re-analysing the data of Halvorsen & Andersen (1984) on cestode infections in Norwegian arctic charr and by analysing new data on nematode infections in Caribbean Anolis lizards.