• Breed group;
  • Colic;
  • Cox proportional hazards regression;
  • Database;
  • Equine;
  • Insurance;
  • Kaplan-Meier;
  • Longitudinal study;
  • Sex


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

Background: Colic is an important cause of morbidity and mortality in horses. In Sweden, an insurance database with diagnostic medical information is maintained on >30% of the nation's horse population.

Hypothesis: The objective was to describe the occurrence of colic, defined by costly veterinary care and life claims, in horses at 1 insurance company during 1997–2002.

Horses: All horses (<21 years of age) with complete insurance for veterinary care and life during the period 1997–2002 were included.

Methods: Colic was defined as conditions where the main clinical sign was abdominal pain and the problem was related to the gastrointestinal system. The analyses included measures of incidence by sex, breed group, age categories, geographical location (urban/other), survival to and survival after colic, medical cost for colic, and multivariable modeling of risk factors related to the event of colic.

Results: In all, 116,288 horses contributed to 341,564 horse years at risk (HYAR). There were 3,100 horses with a colic diagnosis, of which 27% were settled for life insurance. The median gross cost for veterinary care was 4,729 Swedish Kronor (SEK). The overall occurrence and mortality rate of colic was 91 and 24 events per 10,000 HYAR. Survival after colic at 1 month was 76% (95% confidence interval: 75–78%).

Conclusions and Clinical Importance: The occurrence of colic varied with breed group, age, and season. The mortality rates probably reflected the true mortality of colic. The veterinary care rates most likely underestimated of the risk colic because they represent relatively costly events.

Colic is considered one of the most important equine health problems encountered by veterinarians as well as owners1 and has been found to be third after “old age” and “injury wounds/trauma” as the direct cause of death in equids older than 6 months.2

Numerous studies have examined the incidence of and the risk factors for colic.3–7 Survival after colic has been reported less frequently and generally has involved horses undergoing surgery8–13 and less often related the deaths to all colic cases seen in a practice.14 Further, the direct economic cost of colic in the horse has rarely been described. To our knowledge, only 4 studies have presented the direct costs related to colic events.6,15–17

Animal insurance data have been used to describe the incidence of colic in Japan.18 In Sweden, a large insurance database, including information on >30% of the nation's horse population, is maintained by 1 insurance company (Agriaa). Horses can be insured for veterinary care costs, life, or both. Several types of insurance exist, with complete coverage reimbursing the owner for most reasons. General rates of disease occurrence from this database have been presented, including overall incidence of digestive problems.19–21 More detailed analysis of horses with colic is needed to define the potential risk factors for this clinically important disease.

The main objective of this study was to describe the occurrence of colic, as defined by costly veterinary care (ie, exceeding the deductible) or life claims because of colic, in horses with complete insurance for veterinary care and life at Agria during 1997–2002. This included measures of incidence by sex, breed group, across age categories, survival to and survival after a colic event, veterinary care costs for colic, and multivariable modeling of risk factors for the 1st veterinary care event for colic.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

The Insurance Database

There are 2 general types of insurance for horses offered by Agria, namely veterinary care and life insurance. Veterinary care insurance has no age limit and reimburses the owner for most of the expenses if the horse receives costly veterinary care. With life insurance, the owner is reimbursed if the horse dies or is euthanized. Life insurance terminates at 24 or 26 years of age (depending on the breed). Almost all horses are insured for both veterinary care and life.

The insurance process has been described in detail.19–21 Within the 2 general insurance types, there are several subtypes. A complete insurance for veterinary care reimburses the owner for costs incurred if the horse is injured or becomes ill for various reasons. Costs reimbursed include examination and treatment delivered by a veterinarian as well as medications used at the veterinary visit. Complete insurance does not cover travel costs of the veterinarian or costs for prescribed medications. In 2000, the yearly fee per horse for a complete insurance for veterinary care varied between 275 Swedish Kronor (SEK) and 950 SEK (on average, from 1997 to 2002, 1 SEK was approximately US$0.113). The yearly fee for veterinary care insurance is not related to the value of the horse but instead depends on the horse's premium group (for further details, see Egenvall et al19). The deductible for a claim period in 2000 was approximately 1,098 SEK, with the owner paying 20% of veterinary costs that exceeded the deductible. Moreover, the maximal sum of reimbursement each year was approximately 25,000 SEK, above which expenses were fully paid by the owner. Using the insurance does not influence coverage or fees. When the veterinary care insurance is settled, the information on principal diagnosis as well as the net and gross costs from the receipts for veterinary care is entered into the insurance database by the clerk processing the claim.

A complete life insurance policy reimburses the owner if the horse dies, if it is euthanized because of severe injury or illness as judged by a veterinarian, or if use of the horse is permanently impaired because of injury or disease. Whether the horse dies or undergoes euthanasia cannot be determined from the database. The cost for complete life insurance depends on the value of the animal. For example, a warmblood horse life-insured in 2000 for a sum of 30,000 SEK had an annual premium of 2,965 SEK.

Data Management

Data were extracted from the computerized insurance database, as described previously.19,21 The variables used in the present study were sex, date of birth, breed, and geographical location (ie, postal code of owner), name of the horse, information on whether the horse was covered for veterinary care, life or both, as well as diagnostic codes from each receipt, dates of visits to veterinarians, dates of death or euthanasia, veterinary care costs for the claims (gross cost in SEK including the deductible and costs on veterinary care receipts that are not covered by insurance), and data on the life-insurance value for which the horses were life-insured. Dates when horses entered or left the insurance program and reasons for discontinuing were also recorded.

Breeds were combined into 7 groups (warmblooded riding horses, nonracing thoroughbreds, nonracing standardbreds, ponies, Icelandic horses, nonracing coldbloods, and undefined). Age categories were then constructed based on the horse's age as of January 1 (eg, 0< 1 years, 1 < 2 years). Data on horses in age categories 0 < 1 to 20 < 21 were included. Life-insurance value was categorized into 4 levels: <9,000 SEK, 9,000 to <15,000 SEK, 15,000 to <25,000 SEK, and ≥25,000 SEK based on quartiles of the data.

The owner's postal code was used to identify geographically where the horses were housed, with geographic regions defined as southern, middle, or northern Sweden.19 The variable “urban” (horses with owners' postal codes in regions close to the 3 major cities) or “other” was used to signify whether horses lived in an area of higher (higher) or lower (other) human population density. Geographical variables were mainly used to control for potential confounding. The 1st date of veterinary care or life-insurance claim for colic was used to determine the monthly pattern.

Diagnostic Classification

Since January 1, 1995, a hierarchically ordered diagnostic registry for the horse, dog, and cat22 has been used at Agria to assign diagnostic codes correlating the problem or problems of the horse to each receipt. This registry contains approximately 8,000 alphanumeric codes. At the insurance company, each receipt is assigned only 1 diagnostic code, which is derived from the diagnosis provided by the attending veterinarian. Details of the diagnostic registry have been described.20

Study Population

The study population included only those horses with complete insurance both for veterinary care and for life during the period 1997–2002. Horses with different types of insurance coverage during time were excluded (5% of the original database). The time at risk and the claims before 21 years of age were used. Horses that were settled for life insurance because of impaired use (n = 1,013) without actually dying or being euthanized were omitted.21 Those horses for which the postal code could not be matched to the geographical variables were also omitted (<1% of the original database).

Case Definition

In general, an individual was considered to be an incident-case of colic at the time of the first recorded diagnosis for an event that was reimbursed by the veterinary care insurance or settled for life insurance because of colic. Colic was defined as diagnoses for which the main sign was abdominal pain and the problem was related to the gastrointestinal system or the peritoneum, excluding diseases in the reproductive and urinary tract. Diagnoses from the receipts were grouped into 14 categories: nonspecific, impaction, torsion or volvulus, enteritis, sand impaction, incarceration, rupture, invagination, peritonitis, stomach-related diagnoses, strangulating lipoma, and others (eg, foreign bodies, intestinal stricture, and miscellaneous).


Descriptive statistics are presented on the number of horses in the analysis, the number with claims for colic, the distribution of veterinary receipts in those settled for at least 1 colic event, and the occurrence of multiple colic events. Multiple events were arbitrarily defined such that horses were considered to have experienced a new event of colic if receipts were separated by more than 30 days. The age distributions for colic cases are calculated by various categories.

Incidence rates were calculated with exact denominators. Each animal contributed to the denominator the exact time they were at risk (horse years at risk, HYAR) in the database. The incidence rates were multiplied by 10,000 to be interpreted as events per 10,000 HYAR. Although the 1st occurrence of costly colic in this database may not have been 1st colic event in the horse's life, it is highly probable that all such events after insurance enrollment would be captured. Therefore, we use the term incidence rates with this caution.

Incidence rate calculations were carried out for the overall withdrawal rate (based on all horses that left the database, including those that died), overall mortality rate (with any diagnosis assigned, excluding horses that were dead but not claimed), and colic claims (with respect to the 1st registered colic event and mortality rates because of colic). Standard errors (SE) for incidence rates were constructed taking the root of the number of cases and dividing by the HYAR,23 and then multiplying by 10,000. The proportion of cases with a colic life insurance diagnosis of all horses with colic was determined. Breed group-specific incidence rates related to the first colic diagnosis were constructed. Incidence rates related to a colic diagnosis were constructed stratified by geographical location and urban or other.

The crude and breed group-specific proportions of horses that were claimed for colic up to certain ages were calculated for the breed groups with >200 horses with colic, constructing the baseline survival from Cox regression (and without independent variables). Each horse was entered January 1, 1997 or later at the date of enrollment and removed either when it had the 1st registered colic event or when the horse was no longer insured. In the age-related analysis, horses were entered and removed at the actual age. Horses not removed before December 31, 2002 were considered withdrawn (censored) at that date. Ninety-five percent confidence intervals (95% CIs) were calculated for stated proportions.

Age-specific rates for the 1st registered colic event and colic mortality were constructed using the SMOOTH macro,24 which computes age-specific hazards (densities or more general rates) from the baseline survival function computed by PHREG in SAS.b This provides a smoothed estimate of the hazard curve using a kernel smoothing method. The WIDTH parameter was set to one-fifth or one-tenth of the range of event times.

A multivariable Cox regression was developed using as covariates the breed groups (with Icelandic horses as baseline [BL]), sex (BL = stallion), geographical location (BL = north), and the variable urban or other (BL = other). Similarly, horses were entered and removed at actual ages to adjust for any age effect on the outcome. The outcome was time to the 1st colic event. Life-insurance value was not linearly related to outcome. Therefore, as described above, categories were used with <9,000 SEK as BL. All possible 2-way interactions were tested between the variables that remained in the model after backward main effect reduction based on the type 3 criterion. The proportional hazards assumption was investigated by plotting the natural logarithm of the cumulative hazard stratified by each included covariate (log–log plots; from Cox regressions without covariates as described above) against the log of HYAR. Interactions were added at P < .20. A P-value of < .05 was considered significant in the final model. Model fit was inspected using plots of Martingale and deviance residuals against time and against covariates, respectively. CIs presented for hazard ratios (HRs) are 95%. The SASb procedure TPHREG was used for Cox-proportional hazard regression. Survival in horses after the 1st registered colic diagnosis is reported based on Kaplan–Meier analysis by the SAS procedure LIFETEST.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

Descriptive Statistics and Rates

During 1997–2002, 116,288 horses contributed a total denominator of 341,564 HYAR to the 1st registered colic event (Table 1) (346,600 HYAR for mortality). At the end of the study period (December 31, 2002), 61,904 horses were censored and still alive. The number of years each horse contributed ranged from <0.01 to 6 years, with 2.5 years as the median (mean, 2.9 years). Overall, 13,779 horses died with an assigned diagnosis, and the overall mortality rate was 398 deaths (SE 3) per 10,000 HYAR. The 10th, 50th, and 90th percentiles for the life-insurance value were 5,000, 15,000, and 35,300 SEK, respectively, with the maximum value >1,100,000 SEK.

Table 1.   The diagnostic categories of diagnoses for colic events (n = 4,312 and 837 for veterinary care [receipts] and life [deaths], respectively).
DiagnosisVeterinary CareLife Claim
No. of Receipts%No. of Deaths%
  • a

    Others: including intestinal stricture (8 for veterinary care /2 for life), foreign body (4/1), and miscellaneous (3/3).

  • The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years.

“Colic” or “nonspecific” colic diagnosis2,9406855066
Sand impaction78251
Stomach-related diagnoses21< 171
Strangulating lipoma16< 171
Othersa15< 160

There were 3,100 horses with a colic diagnosis settled for either veterinary care or life insurance. Of those, 566 horses were settled for both veterinary care and life insurance. The median age distribution of all the colic cases was 9.3 years with 10th and 90th percentiles of 2.4 and 17.8 years (mean, 9.7 years). By category (for breed, sex, age, life-insurance value, geographical location, and urban or other) all but 2 median ages varied between 8.0 and 10.7 years. The exceptions were stallions (median age, 3.6 years) and Standardbreds (median age, 12.6 years). The 10th, 50th, and 90th percentiles for life-insurance value in the colic cases were 6,000, 17,200, and 27,500 SEK, respectively.

In total, 2,841 horses were settled for veterinary care insurance for colic. There were 4,312 veterinary care receipts, with numbers per horse varying from 1 to 10 (median, 1). There were 1,003 horses with multiple claims with different claim dates, including both veterinary care claims and life claims, of which 368 had receipts more than 30 days apart (12% of the 3,100 horses) and 635 had receipts within 30 days of the initial incident. In the 368 horses, the median of the largest time span between receipts for colic was 0.6 years, with 10th and 90th percentiles of 43 days and 2.9 years, respectively. The distribution of receipts over the diagnostic categories is presented in Table 1 (some have been amalgamated; see table footnote).

The overall rate of colic was 91 events per 10,000 HYAR (SE, 2). Table 2 shows the rate by different categories. Figure 1a shows the age-specific rate of colic. The yearly rates of colic varied between 90 events per 10,000 HYAR (year 1998/2001) and 107 events per 10,000 HYAR (year 1997), with no clear pattern across years.

Table 2.   Details on the incidence rates and age distribution of the colic cases, related to the first registered event.
VariableHorsesHYARRateaSEMRbSEProportion Deadc (%)Median Age at Colic Death
  • a

    Rate-incidence rate for 1st registered event of colic (veterinary care and/or life insurance event) during the time period, expressed per 10,000 HYAR.

  • b

    b MR-mortality rate for colic, expressed per 10,000 HYAR.

  • c

    Number dead with colic/number with colic.

  • The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years.

  • HYAR, horse years at risk; SE, standard error; SEK, Swedish Kronor.

Breed group
 Icelandic horses8,45025,5465751322310.5
Life-insurance value (SEK)
 < 9,000 SEK29,08982,0637032723916.0
 9,000 < 15,000 SEK27,90784,1637932322910.7
 15,000 < 25,000 SEK28,59888,8839832322310.5
 ≥25,000 SEK30,69486,45511542322111.1

Figure 1.  The age-specific rate (per 10,000 horse years at risk, HYAR) of the 1st event of registered colic (morbidity, a) and the age-specific mortality rate of colic (b), developed using Cox regression without covariates. The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years.

Download figure to PowerPoint

There were 825 horses settled for life insurance with a colic diagnosis. Accordingly, the proportion of the colic cases that died with colic as a diagnosis was 27%. This proportion is shown for different categories in Table 2. The 10th, 50th, and 90th percentiles of the age distribution of the colic deaths were 3.0, 12.0, and 19.4 years (mean, 11.7 years), respectively (Table 2).

The mortality rate attributable to colic was 24 deaths per 10,000 HYAR (SE, 1). Figure 1b shows the mortality by age. Table 2 demonstrates the rates in relation to signalment, life-insurance value, and location.

The distribution of both the number of 1st registered claims for colic, as well as the colic deaths by month, is shown in Figure 2. The distribution by different variables for the cost for the colic claims can be found in Table 3. The median cost for veterinary care per horse (including all events) was 4,729 SEK, with minimum and maximum values of 1 and 71,620 SEK.


Figure 2.  The proportion of 1st registered veterinary claims for colic by calendar month (the actual date when the horse was first seen for colic, n = 2841) with 95% (upper) confidence intervals and the date of death for colic deaths (n = 825). The study population is horses in Sweden with complete insurance at a Swedish insurance company during 1997–2002.

Download figure to PowerPoint

Table 3.   Details on the gross cost of veterinary care per horse with colic (in 1,000 SEK).
  • a

    The number of horses add to >2,841 because horses can have claims in different years.

  • The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years.

  • SEK, Swedish Kronor.

Breed group
 Icelandic horses1336.
Life insurance
 < 9,000 SEK4776.
 9,000 < 15,000 SEK6097.
 15,000 < 25,000 SEK8068.

Survival and Multivariable Analysis

Figure 3 illustrates survival to colic (ie, the age when horses first develop colic given no previous colic) by breed group in warmbloods, Thoroughbreds, and ponies (breed groups with >200 cases). When compared with both warmbloods and Thoroughbreds as judged by 95% CIs (data not shown), a significantly smaller proportion of ponies had developed costly colic after approximately 7 years of age.


Figure 3.  The survival to 1st registered colic event by breed group, developed using Cox regression without covariates. The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years.

Download figure to PowerPoint

Survival after colic to colic death at 1 and 6 months was 76.2% (95% CI 74.7–77.7) and 74.9% (95% CI 73.4–76.5), respectively. Overall survival (ie, any cause of death) was just slightly lower (at 6 months, it was 74.7%, 95% CI 73.2–76.3).

There were 116,288 observations and 3,100 events in the multivariable analysis. Except for sex, all tested main effects contributed to the model (Table 4); however, no interactions were significant. The type 3 P-values were <.0001 for all variables, except for urban or other (P= .047). Plots of deviance and Martingale residuals were satisfactory.

Table 4.   Hazard ratios (HRs) with 95% confidence intervals (95% CIs) from Cox regression of time to the 1st veterinary care event with the diagnosis colic.
VariableType 3 P-ValueCategoryHazard Ratio95% CI
  1. The population includes horses with complete insurance at a Swedish insurance company during 1997–2002, excluding ages >21 years. The model includes data from 116,288 horses with 3,100 cases of colic.

  2. BL, baseline; SEK, Swedish kroner.

Region<.0001North (BL)1
 Middle1.00.9, 1.2
 South1.31.1, 1.5
Urban/other.05Other (BL)1
 Urban1.11.0, 1.2
Breed<.0001Icelandic horses-(BL)1
 Warmblooded horses2.11.8, 2.5
 Thoroughbred horses2.11.7, 2.6
 Standardsbreds1.51.2, 2.0
 Coldbloods1.41.1, 1.8
 Ponies1.41.2, 1.7
 Undefined breed1.41.1, 1.9
Life-insurance value<.0001< 9,000 SEK (BL)1
  9,000 < 15,000 SEK1.21.0, 1.3
 15,000 < 25,000 SEK1.41.3, 1.6
 ≥25,000 SEK1.71.5, 1.9


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

The Population and Database

The present study includes horses with complete insurance at 1 insurance company in Sweden. We estimated previously that this database includes a substantial proportion (approximately 34% of the riding horses) of the horse population in Sweden, even though actively training race horses are not included to the same extent, the latter being a major consideration when comparing our findings with many other published colic studies.19 However, the database is maintained as a management tool for the insurance company and not primarily for research. This has several drawbacks, the largest likely being that for each claim only 1 diagnosis can be recorded in the database. Fortunately, this limitation is probably minimal for this study because the colic event will in most cases be the major problem and hence will decrease the risk of missing case events. Another problem is that nonspecific diagnoses are often used when a more specific diagnosis exists.25 Moreover, the database does not facilitate identifying colic cases that underwent surgery. The gross cost for the treatment may be suggestive of surgery, but the lack of specific information makes inferences of medically versus surgically treated colic impossible without contacting the owners or reviewing clinical records, which would be time and labor intensive because of the high number of cases. Another drawback with insurance data is that there is no readily obtainable information on management and usage of the horse. Future research on this topic preferably should include such information obtained from the owners retrospectively or prospectively.

Analytical Issues and Issues with Interpretation

It is likely that some of the horses that developed colic in the present study had colic events previously (especially older horses), outside of the study period, at a low cost during the study period or as a self-resolving episode. Accordingly, the calculated figures of survival after a colic event (ie, proportions of the population that have had colic at various ages) are admittedly conservative. However, we suggest that within this data set it is reasonable to compare survival among variable categories (eg, among breed groups).

As a consequence of the insurance terms, the fact that the horse has insurance coverage most likely influences mortality to some extent. For example, the maximum cost reimbursed per year may influence mortality attributable to colic (as well as other diseases) because the horse might be euthanized because of financial considerations (when the maximum reimbursement has been given for that year). Thus, relatively few horses may undergo a 2nd surgery or receive expensive medical treatment for relapse or recurrence (A. Egenvall, unpublished data). In the present data, economics cannot be separated from purely medical decisions for euthanasia of the horses. The insurance company is often contacted for approval before proceeding with colic surgery, and the value of the horse (life-insurance value) will therefore most likely affect mortality attributable to colic to some degree. Sometimes, it might be more cost-effective for the insurance company to pay out the life insurance rather than to proceed with expensive veterinary treatment if the life-insurance value is low.

Morbidity and Survival to Colic

The incidence of colic has been reported to vary between 3.5 and 18.6 colic episodes per 100 horses and year.4,6,8,26–28 Our estimate of 91 per 10,000 HYAR corresponds roughly to approximately 0.9%, likely because only colic events requiring costs more than the deductible were included.

For the demographic variables, the results in the literature are conflicting. For example, Arabs have been identified as having higher29 as well as lower7 odds ratios for colic compared with other breeds. In the present study, warmbloods and Thoroughbred horses generally had higher HRs for colic than standardbreds, coldbloods, and ponies. However, the latter had higher HRs than Icelandic horses. One reason for the low occurrence in Icelandic horses might be that they are not trained for use until 5 years of age and possibly subjected to a more natural diet with relatively little grain. Had data been available, management factors such as proportions roughage and grain, turn-out practices, and training regimen may have explained some of the observed breed differences. In 2 studies, dietary changes, especially change of hay, contributed to increased risks of colic.30–31

Colic and death attributable to colic increase slowly and fairly constantly through most age categories, increasing more steeply after 16 years of age. Our conclusion is that age is a risk factor. However, reports in the literature have not been consistent. Horses <6 months and <2 years have been reported to be at less risk for developing nonspecific colic5–6 and older horses have a higher risk of colic.26 However, age was not found to be significantly related to colic in a case-control study.31 It is difficult to compare studies because the inclusion criteria, breed composition, and method of information retrieval vary across studies. Moreover, there may be true geographic differences. Therefore, it is essential to conduct regional studies on disease occurrence to more accurately describe the disease pattern in the region of interest.

In 1 study, being a gelding was a protective factor for colic.26 Others have reported that geldings had increased risk for chronic intermittent colic.32 However, the results of the present study are in keeping with other studies6,29–30 that found no association of sex with the risk of colic.

The incidence of colic was somewhat higher in horses insured for a high life-insurance value, whereas colic mortality was not. This may be related to owners of valuable horses with colic signs being more likely to contact a veterinarian at earlier mild stages than those having relatively inexpensive horses. The variable was included because life-insurance value has been shown to be a substantive determinant of usage of both overall veterinary care and life insurance.33–34

Colic Mortality and Survival after Colic

Compared with the morbidity rate where only cases of costly colic are captured, the mortality figure, although not separating deaths from euthanized horses, most likely reflects the true colic mortality rate. However, the proportion of horses that die (case fatality) in our analyses is not identical to death solely because of colic. A few horses are euthanized and life insurance is paid out when the colic may be correctable with surgery, but the cost of surgery exceeds the life-insurance value (and the horse may not have sentimental value or other considerations are taken). This may be 1 reason why the proportion dead was larger than that found in other studies. However, another major reason may be that there were most likely many cases of mild colic that had a favorable prognosis in our horse population that would have had costs below the deductible level for insurance, and thus may not have been captured in our data set to be included as cases. Other studies have found case fatalities for undefined colic of 6.7–15.6%.4,26

Using the proportion of dead horses from Table 2, the highest proportions were for standardbreds, coldbloods, horses with a low insurance value, and horses from the north. Standardbreds and coldbloods often have a lower insurance value, whereas horses in the north generally have longer distances to animal hospitals, which may a priori make the prognosis more guarded. Also, the average age of the standardbreds was higher than that of other breeds, as reflected by the age at colic. Owners of animals with a relatively low insurance value may select euthanasia before surgery. This may explain why the breed group-specific mortality rates are similar, whereas the breed group-specific incidences diverge because of the varying degrees in the proportion dead.

The pattern of survival in the present study was very similar to that reported in another study that also followed both surgically and medically treated horses.14 In both studies, the long-term survival flattened out to 70–75% and remained stable through time.

Geographic region did not influence the occurrence of colic in a study in the continental United States.6 However, in the present study there were small changes such that the south region had somewhat higher HRs of colic compared with the middle and northern areas. Some general differences among the areas relevant to the southern region of Sweden include a progressively higher human population density toward the south35; most of the warmblood colts are registered and raised in southern Sweden (K. Martinsson, Swedish Warmblood Association, personal communication), and pasture areas there are generally smaller. Furthermore, when considering veterinary care of colic cases, owners in the southern region also have excellent access to veterinarians and are more likely to be in closer proximity to veterinary hospitals. Cost was lowest in the north of the country, where the distances to veterinary hospitals are long. The biological reasons behind these associations need further investigation using primary data. Geographic differences have been discussed for some forms of colic (eg, enterolithiasis36 and sand colic in Sweden, from the Agria insurance database, data not shown).

Hillyer et al27 found peaks in spring and autumn, whereas others6 also observed more colic events in spring. In Japan, most cases of acute abdomen were found from May to August, coinciding with feeding fresh fodder.18 Archer et al37 modeled the seasonality of different types of colic, using a time-series technique, and found a pattern in which increased colic coincided with times of management change or periods when horses were more likely to be intensively managed. In the present study, a larger proportion of colic cases was seen in the winter months (October–March), which may be related to a different exercise or feeding regimen during that period, possibly because of harsh winters in Sweden.


In the present study, colic events that exceeded the deductible were captured, which would have excluded mild colic cases and spontaneous recoveries. The mean cost for veterinary services, drugs, and additional care has been estimated at US$160/colic event based on operator-reported colic events.6 This value is considerably lower than the mean cost per colic event found in the present study (8,000 SEK approximately US$904). The main reason for this is most likely the different inclusion criteria in the present study, which included only colic cases claimed and reimbursed at the insurance company and thus always included veterinary visit(s). However, the mean cost for horses undergoing surgery has been estimated at US$3,872 per colic event,6 which, in contrast, is much higher than our figures. Using data from 251 surgically treated horses, an almost 2-fold higher cost for cases with postoperative ileus was found compared with those without.16 The overall median cost for colic per horse per month was US$4.42 in a prospective 1-year study in the early 1990s.15 The overall mean cost in the present study could be estimated using the results from the present study to 195 SEK per horse (8,000 SEK × 2,841 horses with veterinary costs for colic/116,288 horses in the population) or approximately US$22 (195 SEK × US$0.113/1 SEK).

The trend that the cost per colic case has increased over time may be because of different methods of treating colic (eg, more or other medications or earlier surgery) or simply because of increasing costs for fees and medications. The present study cannot answer these questions. However, surgically treated colic has shown a rather dramatic increase in positive prognosis during the later decades. This in itself raises many unanswered questions (eg, related to case selection for surgery).


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

The insurance database produced relevant statistics concerning costly colic in horses on a more general level; however, extrapolation or comparison with other populations should be carried out with caution because of specific limitations of the material. Of the factors studied (eg, breed group, sex, age, season, year, and geographical variables), the occurrence of colic varied most with the breed group, age, and season. The mortality rates probably reflect the true mortality of colic (including euthanasia) in this population. The specific diagnoses registered were not considered reliable for the interpretation of the true occurrence of various types of colic. The incidence of colic in Sweden should be studied using primary data collection in prospective longitudinal studies to achieve a fuller picture of the problem.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

aAgria Insurance, Stockholm, Sweden

bSAS Institute Inc, Cary, NC


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References

This work has been supported by grants from the Foundation for Research, Agria Insurance.


  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Footnotes
  8. Acknowledgment
  9. References
  • 1
    Mellor DJ, Love AS, Gettinby G, et al. Demographic characteristics of the equine population of northern Britain. Vet Rec 1999;145:299304.
  • 2
    NAHMS. 2005 trends in equine mortality 1998–2005. U.S. Department of Agriculture. Available at: Accessed February 8, 2008.
  • 3
    Uhlinger C. Investigations into the incidence of field colic. Eq Vet J 1992;13 (Suppl ):1118.
  • 4
    Tinker MK, White NA, Lessard P, et al. Prospective study of equine colic incidence and mortality. Eq Vet J 1997;29:448453.
  • 5
    Tinker MK, White NA, Lessard P, et al. Prospective study of equine colic risk factors. Eq Vet J 1997;29:454458.
  • 6
    Traub-Dargatz JL, Kopral CA, Seitzinger AH, et al. Estimate of the national incidence of and operation-level risk factors for colic among horses in the United States, spring 1998 to spring 1999. J Am Vet Med Assoc 2001;219:6771.
  • 7
    Hillyer MH, Taylor FGR, Proudman CJ, et al. Case control study to identify risk factors for simple colonic obstruction and distension colic in horses. Eq Vet J 2002;34:455463.
  • 8
    Proudman CJ, Edwards GB, Barnes J, et al. Factors affecting long-term survival of horses recovering from surgery of the small intestine. Eq Vet J 2005;37:360365.
  • 9
    Proudman CJ, Edwards GB, Barnes J, et al. Modelling long-term survival of horses following surgery for large intestinal disease. Eq Vet J 2005;37:366370.
  • 10
    Mair TS, Smith LJ. Survival and complication rates in 300 horses undergoing surgical treatment of colic. Part 1: Short-term survival following a single laparotomy. Eq. Vet. J 2005;37:296302.
  • 11
    Mair TS, Smith LJ. Survival and complication rates in 300 horses undergoing surgical treatment of colic. Part 2: Short-term complications. Eq Vet J 2005;37:303309.
  • 12
    Mair TS, Smith LJ. Survival and complication rates in 300 horses undergoing surgical treatment of colic. Part 3: Long-term complications and survival. Eq Vet J 2005;37:310314.
  • 13
    Mair TS, Smith LJ. Survival and complication rates in 300 horses undergoing surgical treatment of colic. Part 4: Early (acute) relaparotomy. Eq Vet J 2005;37:315318.
  • 14
    Van Der Linden MA, Laffont CM, Sloet van Oldruitenborgh-Oosterbaan M. Prognosis in equine medical and surgical colic. J Vet Intern Med 2003;17:343348.
  • 15
    Lloyd JW, Kaneene JB. Economics of health management in the Michigan, USA equine industry. Prev Vet Med 1997;30:18.
  • 16
    Cohen ND, Lester G, Sanchez LC, et al. Evaluation of risk factors associated with development of postoperative ileus in horses. J Am Vet Med Assoc 2004;225:10701078.
  • 17
    Abutarbush SM, Naylor JM. Comparison of surgical versus medical treatment of nephrosplenic entrapment of the large colon in horses: 19 cases (1992–2002). J Am Vet Med Assoc 2005;227:603605.
  • 18
    Higuchi T. A retrospective survey of equine acute abdominal in a breeding region of Japan based on agricultural mutual relief insurance data. J Eq Sci 2006;17:1722.
  • 19
    Egenvall A, Penell JC, Bonnett BN, et al. Morbidity of Swedish horses insured for veterinary care between 1997 to 2000: Variations with age, sex, breed and location. Vet Rec 2005;157:436443.
  • 20
    Penell JC, Egenvall A, Bonnett BN, et al. Specific causes of morbidity among Swedish horses insured for veterinary care between 1997 and 2000. Vet Rec 2005;157:470477.
  • 21
    Egenvall A, Penell JC, Bonnett BN, et al. Mortality of Swedish horses with complete life insurance between 1997 and 2000—Variations with sex, age, breed and diagnosis. Vet Rec 2006;158:397406.
  • 22
    Svenska Djursjukhusföreningen (Swedish Animal Hospital Association). Diagnosregister för Häst, Hund Och Katt (Diagnostic Registry for the Horse, the Dog and the Cat), 1st ed. Taberg: Svenska Djursjukhusföreningen; 1993:1235.
  • 23
    Breslow NE, Day NE. Statistical Methods in Cancer Research. Volume II. The Design and Analysis of Cohort Studies, Ist ed. Lyon: International Agency for Research on Cancer; 1987:59.
  • 24
    Allison P. Survival Analysis using the SAS® System: A Practical Guide, 1st ed. Cary, NC: SAS Institute Inc; 1995:5759.
  • 25
    Penell JC, Egenvall A, Bonnett BN, et al. Validation of computerized Swedish horse insurance data against veterinary clinical records. Prev Vet Med 2007;82:236251.
  • 26
    Kaneene JB, Miller RA, Ross WA, et al. Risk factors for colic in the Michigan (USA) equine population. Prev Vet Med 1997;39:2336.
  • 27
    Hillyer MH, Taylor FG, French NP. A cross-sectional study of colic in horses in Thoroughbred training premises in the British Isles in 1997. Eq Vet J 2001;33:380385.
  • 28
    Mehdi S, Mohammad V. A farm-based prospective study of equine colic incidence and associated risk factors. J Eq Vet Sci 2006;26:171174.
  • 29
    Reeves MJ, Salman MD, Smith G. Risk factors for equine abdominal cases disease (colic): Results from a multi-center case–control study. Prev Vet Med 1996;26:285301.
  • 30
    Cohen ND, Gibbs PG, Woods AM. Dietary and other management factors associated with colic in horses. J Am Vet Med Assoc 1999;215:5360.
  • 31
    Hudson JM, Cohen ND, Gibbs PG, et al. Feeding practices associated with colic in horses. J Am Vet Med Assoc 2001;219:14191425.
  • 32
    Cohen ND, Peloso JG. Risk factors for history of previous colic and for chronic, intermittent colic in a population of horses. J Am Vet Med Assoc 1996;208:697703.
  • 33
    Egenvall A, Bonnett BN, Olson P, et al. The association between costly veterinary care and 5-year survival of Swedish insured warmblooded riding horses. Prev Vet Med 2006;77:122136.
  • 34
    Egenvall A, Bonnett BN, Wattle O, et al. Veterinary-care events and costs over a 5-year follow-up period for warmblooded riding horses with and without previously recorded locomotor problems in Sweden. Prev Vet Med 2007;83:130143.
  • 35
    Statistics Sweden. 2006 Maps, by subject area. Population. Population sizes. Available at: Accessed October 9, 2006.
  • 36
    Hassel DM, Langer DL, Snyder JR, et al. Evaluation of enterelithiasis in equids: 900 cases (1973–1996). J Am Vet Med Assoc 1999;214:233237.
  • 37
    Archer DC, Pinchbeck GL, Proudman CJ, et al. Is equine colic seasonal? Novel application of a model based approach. BMC Vet J 2006;2:27.