Incidence of Photosensitization in Husbandry Animals: A Meta Study on the Effects of Feed Diversity and Feed Choice

1. Introduction

Cases of plant poisoning in animals are not easy to verify by tracing them back to plants because the symptoms are either unspecific [1], subtle [2] or the suspicious plants ingested are simply not visible any more [3]. Moreover, signs of grazing are inconspicuous and thus often overlooked [4]. One prominent exception between plant poisoning diseases is photosensitization. Apart from the well-known outward symptoms of this disease, there are several and abundant plant species such as of the genus Heracleum and Hypericum that are well-known for their photoactive compounds that may accumulate in the skin and interact with sunlight, causing in worst case severe dermatitis [5,6]. Typically, white colored husbandry animals are affected most often by the disease, although except from a few cases in birds, unequivocal symptoms are to be observed in all kind of colored mammals. These striking features along with the popular aetiology of photosensitization and an unambiguous classification system of photosensitization available [7,8] provided appropriate conditions to get a sufficient number of evidence-based cases for our meta-study to understand why farm animals are poisoned by feed plants. One approach is to explore whether the animals could avoid potential toxic plants by selection behavior [9]. For that question, we reviewed case reports of photosensitization from as early as 1900 up to today by addressing two key variables, feed diversity and feed choice. The more floristically diverse a feed stuff is, the higher is the probability that it contains potentially phototoxic plant species, too. However, both variables have been widely discussed with regard to nutritional ecology [10,11,12] and animal welfare [13], however, they are still neglected topics in farming practice [13]. The goal of this meta study was to investigate whether the level of plant diversity in the feed and the extent of feed choice influence the case number of photosensitization.

2. Materials and Methods

2.1. Search for case reports

Electronical databases such as Web of Science, CAB abstracts and VetSearch were browsed to compile case reports of photosensitization between 1900 and 2022. Reviews as publication format served as data base for single case reports. However, if epidemiological data did not allow to separate several similar cases from each other, those case collections counted for one case. Except for birds, individuals of the typical herbivorous livestock species such as sheep or goat were considered, as well as omnivorous farm animals or wild animals such as kangaroo, wombat and llama.

For this meta study, in each paper any spontaneous outbreak or experiment that caused photosensitization counted as one case. Consequently, the total number of cases of the data base analyzed here, was supposed to be higher than the number of papers. If the disease was a secondary finding of studies with another research focus (e.g. [14]), those instances counted for case reports as well. In contrast, experiments with both artificial causes of photosensitization like medications or animals being force-fed were excluded. Publications stating a non-alimentary related background for hepatogenous photosensitization such as liver-migrating endoparasites (e.g. [15]) were not part of this meta study. Moreover, we supplemented the raw data by notes of doubtful evidence which either the authors themselves had mentioned or from our point of view, the authors failed to prove the agent scientifically sufficient.

2.2. Variables of evaluation case reports

The principal variables for evaluation cases of photosensitization included (1) level of feed examination, (2) kind of phototoxic agent, (3) type of photosensitization, (4) animal species, (5) kind and quality of feed, (6) plant species diversity of feed, and (7) possibility of feed choice. Some principal variables consisted of subsidiary variables (sub variables).

2.2.1. Level of feed examination

We materialized possible parameters of the second variable “level of feed examination” by asking whether the authors examined both the feed quality and botanical composition and whether they carried it out (a) completely, (b) partly or (c) did not indicate any kind of feed examination.

In detail, the criteria that the authors examined both criteria completely, were that they (a) inspected the pasture and indicated further plant species than the already identified phototoxic ones, (b) stated that no other phototoxic plants were growing there (c) noted that a phototoxic plant species was found in large quantities (d) counted fungi spores in case of the presence of any (e.g. [16]), (e) analyzed the feed botanically, (f) stated that no phototoxic plants were found, (g) named both main plant species and those with a mass portion of less than 5%.

The criteria that the authors examined the feed’s quality and botanical composition partly, were that they (a) searched for presence or absence of phototoxic plants (b) stated simply that no other phototoxic plants were present and (c) indicated the presence of “many” species with no further information.

2.2.2. Kind of phototoxic agent

Any associated feed-related ingested organisms were relevant as proven or suspected cause of photosensitization, subsumed as “agents”. All taxa identified in this study as agents corresponded to the following five functional groups, reflecting the classification of feed stuff for grazers and browsers, (1) grasses, (2) herbs, (3) woody plants, (4) bacteria and fungi, (5) unknown agents built another group. In cases where the authors failed to prove the kind of agent, we noted whether they have suspected it.

2.2.3. Type of photosensitization

Photosensitization manifests in four types: (1) primary (direct), (2) secondary (hepatogenous), (3) endogenous and (4) idiopathic type [7,8]. The first type, primary photosensitization, occurs when phototoxic components of plants are taken up through ingestion and arrive in the skin unchanged or through direct skin contact [17]. This first type of photosensitization appears within hours of the ingestion or contact [17]. In contrast, the delayed onset of the disease in the following types may make it difficult to identify the trigger. The second type, hepatogenous photosensitization, is a consequence of liver diseases caused through liver-toxic plant components [18] or fungi [19]. The damaged liver can no longer excrete photoactive phylloerythrin, a by-product of bacterial digestion of chlorophyll, through the bile [20]. Instead, it accumulates in blood and skin where it causes a range of typical photosensitization symptoms [2]. The third type, endogenous photosensitization, is a metabolic effect inherited in cattle and therefore not relevant for this study [21]. The fourth, idiopathic type, describes cases where the cause of the disease is not known so far.

For our study, primary, secondary and idiopathic type of photosensitization, were distinguished.

2.2.4. Animal species

Each mammal species noted as case of photosensitization in a publication accounted for one case, regardless whether it was a husbandry or wild animal species. This approach resulted in the first data base, the number of total single cases (n=x cases). In a few of these cases, more than one animal species was concerned. We added these additional cases to our data base, too. This resulted in the second data base, the number of all nominated animal species (n=x nominations).

2.2.5. Kind and quality of feed

The variable “kind of feed” appeared in the data set as the following sub variables, (1) “fresh” in case of ingestion of living plants either provided on pasture or after being recently harvested in a manger, (2) “conserved” in case of hay, silage or pellets, (3) “both” if animals had access to both kinds of feed and (4) “unknown” if no information about the kind of feed was given by the authors.

The variable “quality of feed” considered the hygienic state of the feed and consisted of two sub variables; (1) “normal hygienic quality” and (2) “poor hygienic quality”. In case of normal quality, there were neither direct nor indirect hints on quality deficiencies apparent in the paper. In case of poor quality, the authors stated either (a) Pithomyces chartarum on pasture grass or bacteria and fungi (mould) in conserved feed stuff or (b) moisture in case of hay, the latter according to criteria described in [22].

2.2.6. Plant species diversity of feed

This principal variable comprised five sub variables, (1) high level diversity, (2) supposed high level diversity, (3) low level diversity, (4) supposed low level diversity, (5) no information available.

In detail, the criteria for the sub variable (1), high level diversity, were if authors (a) stated that there were various weeds present (b) found several plants growing on the property or in conserved feed.

The criteria for the sub variable (2), supposed high level diversity, were when authors (a) mentioned certain key words such as “native vegetation”, “fallow” or plants that are indicator plants for a low nitrogen level in the soil or (b) listed at least four plant species without stating that one was dominating.

The criteria for the sub variable (3), low level diversity, were (a) if the pasture or conserved feed consisted of one or two species, only.

The criteria for (4), supposed low level diversity, were (a) if conserved or fresh feed was stated by the researchers to be dominated or massively invaded by one plant (b) if animals received pellets, only.

The criterion for the sub variable (5), no information available, was that authors did not mention any information on botanical diversity of feed.

2.2.7. Possibility of feed choice

For the second principal variable “possibility of feed choice”, we discerned three kinds of sub variables, (1) feed selection was possible (2) feed selection was not possible (3) feed selection was not known.

In detail, the criteria for the sub variable (1), feed selection was possible, were (a) various kinds of feed stuff were given simultaneously or shortly one after the other (b) the pasture or conserved feed contained at least two plant species without one plant dominating.

The criteria for the sub variable (2), feed selection was not possible, were if (a) a pasture was contaminated with fungi such as Pithomyces chartarum, (b) animals received only pellets, (c) the pasture vegetation was completely dried out, even if biodiversity was high.

The criterion for the sub variable (3), possibility for feed selection not known, was when any relevant information was missing.

2.3. Data analysis

Most of the data were analyzed descriptively, counting the presence or absence of defined traits corresponding to the above-mentioned variables. In addition, between the key variables “plant species diversity of feed” and “possibility of feed choice”, Pearson correlation coefficient r was calculated with Microsoft excel^© on basis of the number of animal nominations. The level of significance was extracted from tables of critical values of correlation coefficient in Cann (2004) [23], based on one-sided test. In order to test whether the mean of differences between the number of cases of two distinct groups was different or not, a one-factorial ANOVA has been carried out. In detail, this procedure was applied on the following variables: (1) primary and secondary type of photosensitization, (2) normal and poor feed quality, (3) low and high feed diversity and (4) feed selection possible or not possible.

3. Results

3.1. Data material

Cases of photosensitization have been obtained from 113 publications and included 66 non-peer reviewed articles (58%) and two ones of unknown status (2%) from Europe, the American continent, Australia, Asia, New Zealand and Africa (Table 1). Besides one non-English study [14] and two German written studies [15,16], mainly English written literature contributed to this meta study. Apart from the German written ones, all other non-English written papers (n=20; 19.5 %) were evaluated on the basis of the English abstract. These abstracts were selected because they contained sufficient information as to the variables concerned here.

The publications described 178 cases (Appendix A). 45 of the papers (40%) were peer-reviewed, 66 ones were not peer-reviewed (58%) and in two cases (2%) it was unknown. Unfortunately, almost one fifth of all publications found, dealing with reports of photosensitization, was not published in English and therefore could not be considered.

73.6% of the cases arouse from only three continents: South America (33.79%), Australia (21.3%) and North-America (18.5%). The majority of feed (81.5%) was either completely (43.3%) or at least partly examined (38.2%). Thus, a minority of feed had not been examined (10.1%) or the status left unknown (8.4%). Only in a few cases, there was a list of the most prominent plant species, thus, our criterion “authors analyzed the feed botanically” was met here.

3.2. Phototoxic agents

In the above-mentioned total of 178 cases, 40 phototoxic agents caused photosensitization (Table 2). The 12 most often involved agents represented three size classes of cases (14-15 cases, 10-13 cases, 6-8 cases) and contained two, four and six agents, respectively. Thus, altogether 12 agents accounted for 118 (29+48+41) of 178 cases (66.3%). The most frequent phototoxic agent was signal grass (Brachiaria decumbens, Poaceae), followed by the herbaceous cottonweed (Froelichia humboldtiana; Amaranthaceae) that alone made up for 16.3% of the cases. Together with four plant genus allocated to the second size class (27.0%) these six out of 40 agents were associated with almost 50% (43.3%) of all cases. In contrast, those more rarely involved agents represented one large size class (n=28), however, accounted for only 60 cases (33.7%). Examples of this large size class were the two herbs tansy (Tanacetum vulgare; Asteraceae) and Dutchman’s breeches (Thamnosma texana; Polygonaceae). In 27 of 178 cases (15.2%), the authors suspected an agent, but failed to prove it.

Represented by 24 genus, phototoxic herbs were the most diverse functional group within the agents (Figure 1). In contrast, only three grasses turned out to be agents; apart from the already mentioned genus Brachiaria, these were Panicum and Alopecurus.

3.3. Animal species as well as kind and quality of feed

Five out of the 113 publications (no. 9, 52, 58, 157, 184; s. appendix A) presented cases in which more than one animal species was involved. This resulted in slightly more nominations of animals (n=184) than case reports (n= 178). In total, there were 184 nominations of animals with symptoms of photosensitization (Table 3).

The altogether 12 distinct species of animals comprised two groups. The first group consisted of four herbivorous typical husbandry animals; sheep, horse, cattle and goat. The second group (“other animals”) contained eight species such as omnivorous farm animals (pig) or those species that were rarely kept as husbandry animals (buffalo, donkey mule) as well as wild animals (fallow deer, grey and red kangaroo, hairy-nosed wombat, llama). Sheep and cattle, allocated to the first group, mainly contributed to the total case number (39.7% and 31.5%) while other animal species contributed the least (11%).

Animals received more feed of normal quality (165 individuals, 89.7%) than of poor quality (19 individuals, 10.3%) (p<0.020) (Table 4). 72.5% of the feed was available in fresh state, thus it was either grazed or recently cut and eaten from a manger. In 66.9% of the cases, fresh fodder was of normal quality. Photosensitization went in 89.3% of the cases back to both fresh forage and normal hygienic quality. In other words, only in a minority (10.7%) poor feed quality resulted in photosensitization.

With 117 of 184 nominations of animals (63.6%), herbs clearly represented the majority of phototoxic agents in all animal species, distantly followed by grasses (25 of 184 nominations; 13.6%) (Table 5).

There was a tendency of higher incidence of secondary photosensitization (97 cases, 98 nominations) than of primary photosensitization (61 cases, 66 nominations; p=0.3524) (Table 6 and Table 7). However, in horses, the incidence of primary photosensitization was clearly higher than of secondary photosensitization (20 versus eight nominations). Moreover, in general, the portion of secondary photosensitization associated with feed of poor quality was higher (73.7%) than the portion of secondary photosensitization associated with feed of normal quality (52.2%). In particular, the relation secondary/primary photosensitization referring to feed quality was 52.7% with feed of poor quality and only 16.4% with normal quality (Table 8).

3.4. Feed diversity and feed choice

The introduction of two key variables “feed diversity” and “feed selection” served to survey whether the case number was associated with the level of diversity or selection possibility. Both variables were positively correlated (r=0.88; p<0.005). The incidence of photosensitization with low feed diversity was significantly higher than the incidence with high feed diversity (80 versus 31 cases; Table 9) (83 versus 32 nominations; Table 10) (p<0.050). Moreover, there were more cases of photosensitization without feed selection (114 cases, 119 nominations) than cases with feed selection (14 cases, 15 nominations) p<0.010. In other words, if the animals had access to high-diversity feed instead of low-diversity feed, the incidence of photosensitization was 27.5% smaller (difference between 44.9% “low diversity” and 17.4% “high diversity”; s. Table 9). However, if the animals had the choice, the incidence of photosensitization was even 56.1% smaller (difference between 64.0% “no selection” and 7.9% “selection possible”). A significant portion of cases was of unknown diversity (37.6%) or unknown selection possibility (28.1%).

Compared to the other animal species, horses showed the lowest ratio of low/high diversity (1.5; s. Table 10), but the highest ratio of no/yes selection possibility (10.5). Thus, horses ingested most often feed of high diversity, however, could select the least between plant species or feed stuff.

4.1. Diversity and selection

One can assume that phototoxic agents such as the most frequently involved plants such as Brachiaria and Froelichia as well as the other four members of two largest sizes classes are most potent phototoxic agents. Such degree of potential phototoxic toxicity may mask any effects of diversity or feed choice because the animals cannot evade even a single, short or superficial contact [24]. Indeed, low incidence of photosensitization under high diversity and free-choice shows that even optimal conditions do not guarantee that animals remain automatically unaffected when they encounter phototoxic plants in good quality feed stuff. However, since no control group such as “no photosensitization” exists in case reports, cases of healthy animals exist neither. Thus none of the variables addressed here can cause an incidence of zero.

Nevertheless, significantly higher incidences associated with the respective opposite sub variables show that high-diversity feed and possible feed selection diminish the incidence of photosensitization by roughly 30% (diversity) and 60% (selection). This finding confirms Provenza et al. (1992) who suggested that an animal is capable of avoiding phototoxic plant species [25]. Both conditions probably prevent photosensitization to a far higher extent together than alone.

The assumably synergistic combination of two correlated variables can be demonstrated in horses. This animal species turns out to be an exception in various regard. First, unlike other animals, horses suffer more often from primary photosensitization than from secondary photosensitization. Primary photosensitization goes back to direct contact to the agent via ingestion or physical contact [26] and thus would represent the typical setting to avoid the disease through allowing feed choice. Second, our data reveal that horses are not only exposed to the most diverse feed, but could the least select between plant species or feed stuff compared to all other animal species regarded here.

4.2. Agents and feed stuff

Herbs as class of highest species diversity itself, cause most of the cases, too. In herbs, the chemical diversity of phototoxic agents is higher than in other functional plant groups [27]. In addition, the majority of cases occurred with fresh feed that includes the unimpaired species-specific composition and concentration of secondary plant metabolites. Therefore, fresh plants are more phototoxic than their dried or conserved counterparts [28]. The significantly smaller number of cases of conserved compared to fresh food may be attributed to still present phototoxic compounds that decompose, yet not completely during any procedure of conservation. For example Araya and Ford (1981) showed that 80% of the hypericin content was lost when St. John´s Wort (Hypericum perforatum) was dried in the sun [29]. Ecologically functional features of living plants such as the complete garniture of secondary metabolites explain the higher number of cases of fresh feed compared with the number of cases where conserved feed is involved.

Moreover, exposed to fresh forage, animals have had the optimal chance to avoid phototoxic plants. In this context it is interesting that at least in 33.7% of the cases, agents were involved that seldom had been identified as phototoxic. This demonstrates that although there are only six plants dominating the majority of cases, a numerous variety of phototoxic agents may cause problems of animal health.

The lower incidence with possible feed choice indicates that large herbivorous animals may sense a significant array of phototoxic substances in plants and are able to avoid those plants to a certain degree. For that ability, normal food quality is essential because it may reduce at least the case number of secondary photosensitization. For example, Minervino et al. [30] write: “Secondary hepatogenous photosensitization is the most important form of photosensitization in Brazil with significant losses to the local cattle industry principally in ruminants raised extensively […]”. However, our data reveal that the significant higher incidence of secondary photosensitization applies only under the condition of poor feed quality. This suggests that not only phototoxic plants but liver-toxic bacteria and fungi, too, described for example in [19,31], have caused the higher portion of secondary photosensitization in other studies.

5. Conclusions

We showed that even optimal conditions, such as a high diversity of feed, a possibility of feed choice and forage of normal quality do not guarantee that all animal individuals keep sound when they encounter phototoxic plants. However, the incidence of photosensitization is remarkably reduced by a third, given there is diet diversity, and even halved, given there is the possibility to select within the feed or within different kind of feed stuffs. We conclude that the more choice an animal has regarding its diet, both in respect to plant diversity and feed choice, the less frequent the incidence of photosensitization is – given an acceptable forage quality. Thus, farmers may prevent photosensitization with their animals by allowing them choice, diversity and quality in feed and feed stuffs.