A retrospective study of equine perinatal loss in Canada (2007 to 2020)

Abstract

Objective

This study aims to identify the most common causes of equine perinatal loss up to 7 d of age in Canada.

Animal

Equine.

Procedure

Necropsy reports from 360 equine perinatal loss cases were acquired from provincial veterinary diagnostic labs across Canada. Each case was classified into a basic cause (noninfectious, infectious, or unidentified) of perinatal loss, then further classified into primary and secondary categories for analysis.

Results

Of the basic causes of perinatal loss, noninfectious causes were the most common. Bacterial causes, such as septicemia, were the most common primary diagnosis overall. Actinobacillus was the most commonly identified bacterial species.

Conclusion

This study showed similar results to those of studies conducted in other countries, including having similar etiologic agents identified. The high prevalence of thyroid hyperplasia identified in this study is notable and was not reported in other, similar retrospective studies, despite being reported in locations other than Canada.

Clinical relevance

Perinatal loss can have important economic consequences for horse breeders; thus, identification of the most common causes is of interest to both veterinarians and breeders.

Introduction

Early foal loss, from birth to 7 d of age, has important economic effects on horse breeding operations, with approximately 3 to 8% of thoroughbred foals dying before they reach training age (1–4). To the authors’ knowledge, the incidence of foal death before competitive or training age in other equine industries has not been reported. Breeding operations represent the cornerstone of all horse industries. In the thoroughbred industry, breeders are often economically unprofitable, with lower-value young horses failing to cover their production costs at sale (5,6). Previous economic assessment in this industry has similarly identified that lower-value thoroughbred broodmares must produce a live foal every year for 7 y to be profitable, and that mares of any value were not profitable if they failed to produce a foal in 2 y out of a 7-year period (7). Thus, equine abortions or early foal losses have the potential to appreciably reduce profitability in an industry already experiencing low or potentially negative profit margins.

There have been several retrospective studies assessing the causes of early foal loss, including in the United Kingdom (8–10), Ireland (11), the USA (12), Australia (13), and Italy (14). There have been 2 Canadian studies published on perinatal morbidity and mortality: one study assessing a single farm in Manitoba (15), and an owner survey-based study assessing loss of thoroughbred foals in western Canada (16). However, to the authors’ knowledge, there have been no published studies assessing perinatal foal mortality specifically encompassing multiple regions of Canada. In the 2010 Equine Canada industry survey, 20% of participants identified themselves as breeders, with an average of 13 breeding mares per owner (17). Considering the high percentage of breeders involved in the Canadian equine industry and the potential economic impacts of early foal loss, a review of the most common causes of perinatal loss would benefit the national equine community. The aim of this study was to evaluate the most common causes of early foal loss in cases submitted to provincial diagnostic laboratories across Canada.

Materials and methods

Ten keywords were selected using a word frequency counter on necropsy reports of equine abortion and perinatal loss cases submitted to Prairie Diagnostic Services (PDS; Saskatoon, Saskatchewan) between 2014 and 2020. The keywords identified were “born,” “umbilical,” “abort*,” “ehv,” “fetal,” “fetus,” “foal,” “gestation,” “herpes*,” and “placent*,” with the asterisk representing all possible suffixes.

Enrolled into the study were the major provincial veterinary diagnostic laboratories in British Columbia (Animal Health Centre, Abbotsford); Alberta (Diagnostic Services Unit, Calgary); Saskatchewan (PDS, Saskatoon); Manitoba (Veterinary Diagnostic Services Laboratory, Winnipeg); Ontario (Animal Health Laboratory, Guelph); Quebec (Centre de Diagnostic Vétérinaire de l’Université de Montréal, Saint-Hyacinthe); New Brunswick (Veterinary Laboratory Services, Fredericton); Nova Scotia (Animal Health Lab, Truro); Newfoundland and Labrador (Animal Health Laboratory, St. John’s); and Prince Edward Island (Atlantic Veterinary College, Charlottetown). Nova Scotia and Newfoundland and Labrador were unable to provide case data. For the remaining participating laboratories, the keywords identified were used to retrieve cases of equine abortion and perinatal loss for the years 2007 to 2020. If the laboratory was unable to provide cases for the entire time period, the longest time period ending in 2020 was queried. When available, the case history, full necropsy, histology reports, and the results of any ancillary testing were retrieved for each case. Any identifying information was removed from the reports before submission to the study.

Each report submitted was identified as either an abortion/stillbirth case or a perinatal loss case based on whether the lungs had inflated, as identified in the necropsy report. The reports were reviewed by a veterinary pathology graduate student to categorize the cases into a basic cause of loss (infectious, noninfectious, or unidentified), as well as primary and secondary diagnosis categories (Table 1). Some reports did not provide sufficient detail to assign the case to a basic cause category; these cases were listed as “uncategorized”. Cases for which there was no final diagnosis were categorized as far as possible, with the final diagnosis listed as “unknown”. Any diagnosis categories that represented less than 10% of the cases were grouped under “other” for data analysis.

Table 1.

Numbers of equine perinatal loss cases submitted to provincial veterinary diagnostic laboratories by year of case submission, region, case province of origin, and submitting laboratory.

Region	Submission year (2007 to 2020)														Total
	07	08	09	10	11	12	13	14	15	16	17	18	19	20
Pacific	1	6	1	2	6	3	2	0	2	1	2	1	0	0	27
British Columbia (PDS, AHC)	1, —	2, 4	0, 1	0, 2	3, 3	0, 3	0, 2	0, 0	0, 2	0, 1	1,1	0, 1	0, 0	0, 0	27
Prairie	4	11	11	3	6	6	12	7	14	11	11	8	4	4	112
Alberta (PDS, DSU)	0, —	3, —	7, —	0, 1	2, 4	3, 1	4, 4	1, 2	2, 2	0, 4	2, 1	2, 3	0, 0	1, 1	50
Saskatchewan (PDS)	4	8	4	2	0	2	4	4	10	6	8	3	2	2	59
Manitoba (VDSL)	—	—	—	—	—	—	—	—	—	1	0	0	2	0	3
Central	7	22	24	24	21	15	10	16	14	9	15	12	17	9	215
Ontario (AHL)	7	22	24	24	21	15	10	16	14	9	15	12	16	8	213
Quebec (FMV)	—	—	—	0	0	0	0	0	0	0	0	0	1	1	2
Atlantic	—	—	—	0	0	0	1	2	2	0	1	0	0	0	6
New Brunswick (VLS)	—	—	—	—	—	—	0	0	0	0	0	0	0	0	0
Prince Edward Island (AVC)	—	—	—	0	0	0	1	2	2	0	1	0	0	0	6
Total	12	39	36	29	33	24	25	25	32	21	29	21	21	13	360

Cells containing a dash did not have case submission data available for that year.

PDS — Prairie Diagnostic Services, Saskatoon, Saskatchewan; AHC — Animal Health Centre, Abbotsford, British Columbia; DSU — Diagnostic Services Unit, Calgary, Alberta; VDSL — Veterinary Diagnostic Services Laboratory, Winnipeg, Manitoba; AHL — Animal Health Laboratory, Guelph, Ontario; FMV — Faculté de médecine vétérinaire, Saint-Hyacinthe, Quebec; VLS — Veterinary Laboratory Services, Fredericton, New Brunswick; AVC — Atlantic Veterinary College, Charlottetown, Prince Edward Island.

The information gathered from each necropsy report was recorded in Microsoft Excel spreadsheets (2010) as previously described (18). The regions of Canada were identified as Pacific (British Columbia); prairie (Alberta, Saskatchewan, Manitoba); central (Quebec, Ontario); and Atlantic (Prince Edward Island, New Brunswick). The breed of the foal was classified into one of the following categories: donkey, baroque, light pleasure, draft, miniature horse, standardbred, stock horse, thoroughbred, warmblood, pony, and other. Foal ages were identified as Day 0 to 1, which included foals between 0 h and 48 h of age; Day 2 to 3, for foals between 49 and 72 h of age; Day 4 to 5, for foals between 73 and 120 h of age; and Day 6 to 7, for foals between 121 and 168 h of age. Cases of twinning in which both foals were submitted were recorded as separate cases.

The data were imported into the R statistical package (version 4.1.0; The R Foundation, Vienna, Austria) for analysis. Relative proportions of basic causes of foal loss, primary diagnosis categories, secondary diagnoses categories, and various risk factors were calculated. A Fisher’s exact test was used for assessing differences in relative proportion when < 5 groups were assessed, whereas a Pearson’s chi-squared test (X ² ) was used for assessment of > 5 groups. Each test was followed by pairwise comparisons using Fisher’s exact tests, with Holm’s P-value adjustment if required. All statistical analyses were conducted at a 95% confidence level, with a P-value of < 0.05 considered to be statistically significant.

Results

Three hundred and sixty cases were submitted from the 7 veterinary diagnostic labs queried between 2007 and 2020, ranging from no submitted cases (New Brunswick) to 213 cases (Ontario). The results are summarized in Table 1. Cases of perinatal loss were most frequently identified as noninfectious (n = 210, 58.3%), followed by infectious (n = 124, 34.4%) (Figure 1). Cases that could not be classified into one of these categories were listed as unidentified (n = 26, 7.2%). The proportions of basic causes of foal loss did not change significantly by year of submission [Pearson’s X ² (26, N = 360) = 27.568, P = 0.380]. There was no significant difference in the proportions of basic causes of foal loss by foal sex (Fisher’s exact; P = 0.3834) or foal breed category [Pearson’s X ² (20, N = 321) =16.723, P = 0.671].

Figure 1.

Relative frequencies of basic causes of foal loss and primary categories and the most common secondary categories identified for cases of equine perinatal loss up to 7 d of age submitted to provincial veterinary diagnostic laboratories in Canada.

The majority of the noninfectious cases identified had an unknown final diagnosis (n = 54, 25.6%). The most common primary diagnoses within this category were respiratory (n = 37, 17.6%), traumatic (n = 28, 13.3%), congenital (n = 26, 12.4%), and nutritional (n = 23, 11.0%). The remaining primary diagnoses, such as neonatal hypoxia, fetoplacental, and other organ systems, represented < 10% of the cases and were aggregated under “other” (n = 42, 20.0%). The respiratory causes of foal loss were most frequently identified as aspiration pneumonia (n = 24, 64.9%) and pulmonary atelectasis (n = 9, 24.3%), with occasional cases of acute respiratory distress syndrome and 1 case of nonspecific pulmonary edema. The majority of the aspiration pneumonia cases were related to aspiration of amniotic fluid (n = 20, 83.3%), whereas 4 cases were due to postpartum aspiration of milk. The congenital malformations seen were primarily gastrointestinal (n = 11, 42.3%), including cleft palate, atresia ani and atresia coli, diaphragmatic hernias, and intestinal agangliosis (lethal white syndrome). Of the nutritional cases, thyroid hyperplasia represented the vast majority of cases (n = 21, 91.3%), with 2 cases of white muscle disease. Cases of thyroid hyperplasia were reported in the Pacific, prairie, and central regions, whereas white muscle disease was reported only in the central region.

The infectious cases of early foal loss were primarily bacterial, representing 88.7% of the infectious cases (n = 110). Additionally, bacterial infections were the most common primary diagnosis when comparing across all 3 basic causes of foal loss (n = 110, 30.6%; Table 2). There were smaller numbers of viral loss (n = 12, 10.0%), and no losses associated with fungal cases were identified. There were 2 cases with an unidentified etiology that was suspected to be infectious in nature. All of the viral cases were attributed to equine herpesvirus infection (n = 12, 100.0%). The cases with a bacterial cause were primarily cases of septicemia (n = 73, 66.4%), with lower numbers of respiratory (n = 17, 15.5%) and gastrointestinal (n = 14, 12.7%) infections. There were occasional cases attributed to placentitis (n = 4), hepatitis (n = 1), and omphalitis (n = 1).

Table 2.

The most common primary diagnoses identified in a retrospective survey of equine perinatal loss up to 7 d of age submitted to veterinary diagnostic laboratories in Canada from 2007 to 2020, organized into basic cause (noninfectious, infectious, unidentified) and primary diagnosis categories.

Diagnosis categories	Number of cases	Proportion of basic cause	Proportion of primary category	Overall proportion
Basic cause: Noninfectious	210	—	—	—
Unknown	54	25.7%	—	15.0%
Respiratory	37	17.6%	—	11.4%
Aspiration pneumonia	24	—	64.9%	—
Atelectasis	9	—	24.3%	—
Other	4	—	10.8%	—
Traumatic	28	13.3%	—	7.8%
Dystocia-associated	13	—	46.4%	—
Other	15	—	53.6%	—
Congenital	26	12.4%	—	7.2%
Gastrointestinal	11	—	42.3%	—
Other	15	—	57.7%	—
Nutritional	23	11.0%	—	6.4%
Thyroid hyperplasia	21	—	91.3%	—
White muscle disease	2	—	8.7%	—
Other	42	20.0%	—	11.7%
Basic cause: Infectious	124	—	—	—
Bacterial	110	88.7%	—	30.6%
Septicemia	73	—	66.4%	—
Respiratory infections	17	—	15.5%	—
Gastrointestinal	14	—	12.7%	—
Other	6	—	5.5%	—
Viral	12	9.7%	—	3.3%
Equine herpesvirus	12	—	100.0%	—
Unknown	2	1.6%	—	0.6%
Basic cause: Unidentified	26	—	—	—

The most commonly identified secondary categories for selected primary diagnoses are shown indented below each primary category. For each primary category, the proportion of the overarching basic cause represented by that category, and the proportion represented by that category when comparing all submitted cases, regardless of basic cause, is indicated. The proportion of the primary category represented by each secondary diagnosis is also indicated.

The most commonly identified bacterial genera on bacterial culture were Actinobacillus (n = 21, 42%), Clostridium (n = 12, 24%), Streptococcus (n = 6, 12%), and Escherichia (n = 5, 10%). Also identified were cases of Enterococcus, Klebsiella, Pseudomonas, Salmonella, and Staphylococcus. All of the Escherichia cases were attributed to E. coli, and all were diagnosed as septicemia (n = 5, 100.0%). The most commonly identified Actinobacillus species was A. equuli (n = 19, 90.5%), with single cases of A. arthriditis and A. suis. All of the Actinobacillus cases were diagnosed as septicemia (n = 21, 100.0%). The clostridial species identified were primarily C. perfringens, mainly causing gastrointestinal infections and occasional cases of septicemia. There were single cases of C. difficile enterocolitis and a single case of C. piliforme hepatic disease. Streptococcus zooepidemicus was the most commonly identified streptococcal species (n = 3). Other streptococcal species were reported to cause septicemia (n = 3) and respiratory infection (n = 2).

For 14 of the 26 unidentified cases, lesions were noted in the placenta, gastrointestinal tract, respiratory tract, urinary tract, or liver, although no specific cause or etiology was identified. Ancillary testing, such as bacterial culture or virology, was conducted in 9 of the cases, with negative results. The majority of unidentified cases had no final diagnosis (n = 12, 46.2%), with ancillary testing conducted on 6 of the cases. Across all of the noninfectious, infectious, and unidentified cases, there were 68 cases with no final diagnosis reported (18.9%). Two cases were of a suspected, but unidentified, infectious etiology; 54 were suspected to be noninfectious; and 12 cases had lesions that could not be classified as either infectious or noninfectious and were reported as part of the unidentified category.

By foal age

The results of analysis when considering foal age categories is summarized in Table 3. The majority of foals submitted were between 0 and 1 d of age (n = 195, 54.7%), followed by foals from 2 to 3 d of age (n = 87, 24.2%), foals from 4 to 5 d of age (n = 34, 9.4%), and foals from 6 to 7 d of age (n = 19, 5.3%). Twenty-five cases did not have an age of the foal supplied in the history (n = 25, 6.9%). The relative frequency of cases of foal loss declined with each subsequent d of foal age.

Table 3.

The most common primary diagnoses identified in a retrospective survey of equine perinatal loss up to 7 d of age submitted to veterinary diagnostic laboratories in Canada from 2007 to 2020, organized into age of foal, basic cause (noninfectious, infectious, unidentified), and primary diagnosis categories.

Diagnosis categories	Foals 0 to 1 d of age (n = 195)				Foals 2 to 3 d of age (n = 87)
	n	% of age cat.	% of basic cause	Over. % of age	n	% of age cat.	% of basic cause	Over. % of age
Noninfectious	146	74.9%	—	—	33	37.9%	—	—
Unknown	48	—	32.9%	24.6%	1	—	3.0%	1.1%
Respiratory	24	—	16.4%	12.3%	7	—	21.2%	8.0%
Traumatic	20	—	13.7%	10.3%	3	—	9.1%	3.4%
Congenital	15	—	10.3%	7.7%	9	—	27.3%	10.3%
Nutritional	11	—	7.5%	5.6%	5	—	15.2%	5.7%
Other	28	—	19.2%	14.4%	8	—	24.2%	9.2%
Infectious	40	20.5%	—	—	46	52.9%	—	—
Bacterial	34	—	85.0%	17.4%	43	—	93.5%	49.4%
Viral	6	—	15.0%	3.1%	2	—	4.3%	2.3%
Unknown	0	—	0.0%	0.0%	1	—	2.2%	1.1%
Unidentified	9	4.6%	—	—	8	9.2%	—	—
Diagnosis categories	Foals 4 to 5 d of age (n = 34)				Foals 6 to 7 d of age (n = 19)
	n	% of age cat.	% of basic cause	Over. % of age	n	% of age cat.	% of basic cause	Over. % of age
Noninfectious	11	32.4%	—	—	5	26.3%	—	—
Unknown	0	—	0.0%	0.0%	1	—	20.0%	5.3%
Respiratory	3	—	27.3%	8.8%	1	—	20.0%	5.3%
Traumatic	2	—	18.2%	5.9%	1	—	20.0%	5.3%
Congenital	1	—	9.1%	2.9%	0	—	0.0%	0.0%
Nutritional	4	—	36.4%	11.8%	1	—	20.0%	5.3%
Other	1	—	9.1%	2.9%	1	—	20.0%	5.3%
Infectious	21	61.8%	—	—	11	57.9%	—	—
Bacterial	20	—	95.2%	58.8%	9	—	81.8%	47.4%
Viral	1	—	4.8%	2.9%	1	—	9.1%	5.3%
Unknown	0	—	0.0%	0.0%	1	—	9.1%	5.3%
Unidentified	2	5.9%	—	—	3	15.8%	—	—

For each primary category, the proportion of the overarching basic cause represented by that category, and the proportion represented by that category when comparing all submitted cases, regardless of basic cause, is indicated.

Foals from 0 to 1 d of age most frequently had noninfectious causes of death (n = 146, 74.9%). For the remaining age categories, infectious causes of death were more common (Table 3). On a Fisher’s exact test, the proportions of basic causes of foal loss were significantly different based on foal age category. Post-hoc pairwise comparisons using Fisher’s exact tests with Holm’s P-value adjustment showed that foals from 0 to 1 d of age were significantly more likely to have noninfectious causes of death compared to all other age categories (2 to 3 d: P < 0.001, 4 to 5 d: P < 0.001, 6 to 7 d: P < 0.001).

For all 4 age categories, bacterial infections were the most common infectious causes of disease (Table 3). Actinobacillus species were most commonly identified for 0- to 1-day-old (n = 9, 60.0%), 2- to 3-day-old (n = 8, 36.4%), and 4- to 5-day-old (n = 4, 50.0%) foals, with no cases identified in 6- to 7-dayold foals. There was no significant difference in the distribution of the different genera of bacteria between foal age categories [Pearson’s X ² (24, N = 50) = 28.852, P = 0.226]. Foals from 0 to 1 d of age had the highest rate of viral infections across the 4 age categories, at 15.0% (n = 6). All of the viral infections with an identified etiology were attributed to equine herpesvirus. The proportion of bacterial and viral infectious loss was not significantly affected by the age category of the foal on a Fisher’s exact test (P = 0.216).

Respiratory conditions were the most commonly identified noninfectious disorder in foals from 0 to 1 d of age (n = 24, 16.4%), including conditions such as aspiration pneumonia and pulmonary atelectasis. The distribution of cases for the most common noninfectious primary diagnoses was significantly affected by the age category of the foal [Pearson’s X ² (21, N = 210) = 34.538, P = 0.032]. Based on Fisher’s exact tests with Holm’s P-value adjustments, the proportion of unknown cases compared to several other primary diagnoses was higher for 0- to 1-day-old foals than for 2- to 3-day-old foals; however, only the comparison of congenital disease to unknown causes yielded a significant result (P = 0.018).

By region

For the central, Pacific and prairie regions, noninfectious causes of perinatal loss were most common. The central region had the highest rate of noninfectious loss (n = 130, 60.5%), followed by the prairie (n = 63, 56.3%) and Pacific (n = 15, 55.6%) regions. The Pacific region had the highest rate of infectious causes of perinatal loss (central: n = 74, 34.4%; Pacific: n = 10, 37.0%; prairie: n = 37, 33.0%). Of the 6 cases from the Atlantic region, 3 were infectious in nature, 1 was undetermined, and 2 were noninfectious. On a Fisher’s exact test, there was no significant difference in the proportions of basic causes of foal loss between the different regions of Canada (P = 0.352). There was no significant difference in the distribution of identified bacterial genera in infectious disease between the regions [Pearson’s X ² (24, N = 73) = 23.261, P = 0.505], and all viral cases with an identified etiology were attributed to equine herpesvirus. Specific analysis of the regional differences for thyroid hyperplasia was conducted due to its historical identification in western Canada (19). Analysis revealed statistically significant differences in the proportions of thyroid hyperplasia compared to all other causes of perinatal loss on a Fisher’s exact test (P = 0.003). On pairwise comparison using Fisher’s exact tests, with Holm’s P-value adjustments, the relative proportion of thyroid hyperplasia cases was significantly higher in the prairie region than in the central region (P = 0.004).

By submission type

Of the 358 cases that had the submission type included in the report, 306 were whole-body submissions and 52 cases were limited to specimens only. Two hundred and fifty-two of the whole-body submissions (82.4%) and 38 of the specimen submissions (73.1%) received a diagnosis. There was no statistically significant difference in the rate of diagnosis between the 2 submission types on a Fisher’s exact test (P = 0.127). The prairie region had the lowest rate of full-body submissions, at 67.0% (n = 75), compared to the other regions, which ranged from 77.8% (Pacific, n = 21) to 100.0% (Atlantic, n = 6). Histopathology was done in 330 of the 354 cases where the information was available (93.2%). Additional ancillary testing, such as bacteriology or virology, was done in 228 cases (63.3%).

Discussion

The results of this study reveal several important findings regarding perinatal foal loss in Canada. When considering broad categories of foal loss, such as infectious versus noninfectious, noninfectious causes of early foal loss are most common, and include respiratory conditions, traumatic injuries, congenital malformations, and nutritional disorders. Of all the primary diagnosis categories, bacterial infections were the most common when comparing across noninfectious and infectious. The majority of these cases were cases of septicemia. Foals aged 0 to 1 d of age were more commonly diagnosed with noninfectious causes of perinatal loss when compared with older foals. Older foals were most commonly diagnosed with bacterial infections. There were no statistical differences in the broad categories of perinatal loss reported between regions. Specific investigation into thyroid hyperplasia revealed a statistically significant increase in diagnosis in the prairie region, which correlates with its historical diagnosis in this region (19). The type of submission (e.g., necropsy in a jar, full body) did not have a significant effect on the rate of diagnosis for perinatal loss cases.

There have been several studies analyzing the causes and prevalence of early foal loss in other countries (1,8–14); however, the examined age range and methodology of case selection has varied significantly. Although direct comparison between other studies and the present study is thus difficult, the general trends presented by each study may be valuable for comparison with the present study.

There was a higher proportion of noninfectious cases of early foal loss than infectious cases in this study, which parallels findings in the United Kingdom (8,10), Ireland (11), Australia (13), Switzerland (20), and on a single farm in Manitoba (15). The ages assessed in these studies mainly ranged from up to 7 d of age (8) to up to 60 d of age (10), with 1 study extending to up to 1 y of age (11). In the present study, respiratory diagnoses, such as aspiration pneumonia and atelectasis, were the most common noninfectious diagnoses. This finding is most similar to Whitwell’s 1980 study in the United Kingdom, which also showed a high proportion of respiratory diagnoses (8). These diagnoses were not commonly reported in the other studies listed. These studies frequently reported trauma as the most common noninfectious diagnosis, ranging between 44 to 58% of noninfectious cases (10,13). This proportion was higher than in the present study (23%), likely due to an increased prevalence of traumatic injuries in older foals. The Irish study by Galvin and Corley reported the highest rate of congenital abnormalities among these studies: 50% of noninfectious diagnoses (11), which was also higher than the findings in the present study (12%).

Interestingly, the higher proportion of noninfectious cases in this study is dissimilar to the Canadian findings of Morley and Townsend in 1997. In their study, they used owner-submitted surveys to assess foal morbidity and mortality up to 14 d of age in the prairie and Pacific region. From their surveys, they identified a higher proportion of foal loss from infectious causes than from noninfectious causes (16). This finding was not supported by the current study, and in fact, the prairie and Pacific regions had the lowest rates of infectious perinatal loss of the 4 regions. These differences may be accounted for by the slightly longer age period assessed in the 1997 study, allowing more time for infectious disease to establish and subsequently cause mortality. In addition, it is possible that noninfectious cases are more likely to be submitted to a diagnostic laboratory for necropsy (for either diagnosis or confirmation of a suspected diagnosis) compared to infectious cases, which are more readily confirmed clinically and therefore may be less likely to be submitted. These non-submitted infectious cases would be captured in a survey format, which does not require submission of the foal for necropsy. Furthermore, the 10-year difference between the 2 studies likely brought new scientific advancements in preventing and treating infectious causes of perinatal loss, thus reducing the proportion of infectious cases in our later study.

None of the previously conducted retrospective studies discussed reported nutritional causes of perinatal loss, which represented 19% of the noninfectious cases in the present study. The vast majority of the nutritional cases in this study were cases of thyroid hyperplasia attributed to congenital hypothyroidism (n = 21, 91.3%). This condition was first described in western Canada in 1981 (21), and may be linked to pregnant mare diets high in nitrate or low in iodine (19). However, no specific etiology has been identified. This condition was initially considered to be limited to western Canada, making our identification of congenital hypothyroidism in Canada, particularly in the prairie region, unsurprising. However, its absence in other studies is somewhat unexpected. This absence is most likely due to lack of identification of the condition rather than true absence, as there have been increasing reports from around the world, including a recent publication from Finland (21).

Other studies, such as those conducted in Colombia (22), New Zealand (23), Italy (14), and the USA (12), reported a higher proportion of infectious causes of perinatal loss. These studies ranged in age period from up to 7 d (14) to up to 30 d (6,14). The age period assessed for this group of studies was shorter than for the previous group, which had primarily noninfectious cases, so it is unlikely that the increasing risk of infectious diseases with increasing foal age identified in this study accounts for the difference in findings. The higher rate of infectious perinatal loss may be due to differences in data collection between the studies, or differences in management of the mare and foal population examined. This suggestion is potentially supported by the lack of statistically significant differences in proportions of basic causes of early foal loss across many regions of Canada. Although Canada has widely diverse climates, environments, and geographic regions, horses are managed in similar ways across the country (17), suggesting that environment or geography has less of a role in early foal loss than management techniques. The higher prevalence of infectious causes in these countries may also reflect a true difference in disease prevalence; however, this would be difficult to assess without specific studies.

In the present study, bacterial causes of perinatal loss were the most common overall, which is similar to the findings of Sturgill and Carter in Kentucky in 2004 to 2005, where all of the infectious cases were associated with bacterial etiologies. Of these cases, 1/2 were identified as septicemia, with the remaining cases distributed between gastrointestinal and respiratory infections. This is similar to the findings in the present study, where 66% of bacterial cases were septicemic, with 13% and 15% affecting the gastrointestinal and respiratory tracts, respectively. Marenzoni et al, in Italy, had dissimilar findings in 2012, with 73% of their cases being attributed to viral causes; specifically, equine herpesvirus-1. The remaining studies in this group did not identify specific types or etiologies of infectious perinatal loss for comparison.

Of the studies reviewed, E. coli, A. equuli, Klebsiella spp., and Streptococcus spp. were identified in cases of perinatal loss, similar to the findings in the present study. These pathogens are common causes of foal septicemia (24,25), which is consistent with our findings. Actinobacillus suis and A. arthritidis were not specifically identified in the other studies; however, there have been previous reports of septicemia in foals caused by these species (26,27). One study identified a case of C. perfringens (8), a common cause of neonatal diarrhea (28) that was also identified in our study. Only cases of equine herpesvirus were identified in the cases reviewed and in the present study; however, it is important to note that other viruses, such as rotavirus, can affect foals (28).

That foals from birth to 24 h of age were more susceptible to noninfectious causes of death is expected. These foals are more likely to be diagnosed with conditions relating to the birth process, such as dystocia or aspiration pneumonia, as well as congenital malformations. Interestingly, this age category had the highest rate of viral infections, at 15.0% of cases. This finding was not expected, as typically infectious causes of perinatal loss have an incubation period that precludes infection in newborn foals of this age (29). However, all of these cases were attributed to equine herpesvirus, which can infect foals during late gestation. Infected foals typically present with signs of respiratory and liver disease and frequently die within a few days of birth (30). The more surprising finding is the low rate of viral infections in the older age categories, as rotavirus has been identified as the most common cause of foal enteritis in the USA, Ireland, and England (31). Of the cases included in this study, only 1 3-day-old foal was tested for rotavirus, with a negative result. The low level of testing identified in this study may represent a potential gap in diagnostic information that may affect the rate of successful diagnosis.

The noninfectious causes of equine perinatal loss were the most commonly reported in this study. This was similar to a previous study in the United Kingdom. Overall, the causes of neonatal mortality in Canada were similar to causes identified in other countries, including having similar viral and bacterial etiologic agents. The most striking exception was the identification of congenital hypothyroidism in our study, with no cases identified in the other retrospective studies used for comparison. Because this condition has now been diagnosed in other countries, specific training of pathologists in identification of this disease may be warranted in countries where it has not previously been evident. Finally, our study also emphasizes the need for further research into the underlying pathogenesis of this condition, the potential role of pregnant mare nutrition in disease development, and potential prevention measures or treatments.