Risk Factors for Equine Postoperative Ileus and Effectiveness of Prophylactic Lidocaine
Performed at Department of Large Animal Internal Medicine, Faculty of Veterinary Medicine, Ghent University, Ghent, Belgium. Results were presented at the BEVA Congress, Liverpool, September 2008.
Corresponding author: Sara Torfs, Department of Large Animal Internal Medicine, Faculty of Veterinary Medicine, Ghent University, Salisburylaan 133, 9820 Merelbeke, Belgium; e-mail: firstname.lastname@example.org
Background: Postoperative ileus (POI) is a frequent and often fatal complication of colic surgery. Reliably effective treatments are not available.
Objectives: To determine risk factors and protective factors associated with POI, and to assess the effect of lidocaine IV on short-term survival.
Animals: One hundred and twenty-six horses that underwent small intestinal colic surgery and that survived for at least 24 hours postoperatively.
Methods: Retrospective cross-sectional study. The association of 31 pre-, intra-, and postoperative variables with POI and the association of lidocaine treatment with short-term survival were investigated. Associations were evaluated with univariable logistic regression models, followed by multivariable analysis.
Results: Significant associations of high heart rate (odds ratio [OR] = 1.05, 95% confidence interval [CI] 1.03–1.08), the presence of more than 8 L of reflux at admission (OR = 3.02, 95% CI 1.13–8.02) and the performance of a small intestinal resection (OR = 2.46, 95% CI 1.15–5.27) with an increased probability of POI were demonstrated. Prophylactic lidocaine treatment was significantly associated with a reduced incidence of POI (OR = 0.25, 95% CI 0.11–0.56). Lidocaine treatment was also significantly associated with enhanced short-term survival (OR = 0.30, 95% CI 0.09–0.98).
Conclusions and Clinical Importance: The variables associated with an increased risk of POI can be useful in identifying horses at risk of POI and in providing a more accurate prognosis. The results are supportive for lidocaine IV as an effective prokinetic treatment after small intestinal colic surgery.
bis in die (twice a day)
constant rate infusion
capillary refill time
mean arterial pressure
partial arterial oxygen tension
quater in die (4 times a day)
Postoperative ileus (POI) is a serious and often fatal complication in horses with small intestinal colic. Recent studies indicate that the overall prevalence of POI among horses with colic ranges from 10 to 19%, increasing to 50% when only horses with small intestinal colic are taken into account.1–7 Risk factors previously associated with the development of POI include an increased PCV and heart rate, the presence of reflux at admission, small intestinal involvement, increased duration of anesthesia, and performance of a small intestinal resection.2–4,6,8 The associated clinical signs include gastrointestinal reflux, small intestinal distention, hemoconcentration, tachycardia, and abdominal pain.8,9
The success rate of treating ileus is rather low and associated financial costs can be high. POI is a risk factor for repeat celiotomy4,10 and horses that recover from POI are more likely to suffer from recurrent episodes of colic after discharge.11 The high case-fatality rates associated with equine POI, ranging from 14 to 52%,1,6,7,12,13 are in sharp contrast with the human situation, where the condition is rarely fatal.14
By identifying relevant predisposing or protective factors, more insight can be gained in the pathophysiology of POI in horses, which will help to tackle this troublesome problem and to develop optimal, evidence-based treatment protocols. Currently, lidocaine is the prokinetic agent most frequently used for treatment of POI in equine practice,15 although scientific evidence on its prokinetic and analgesic effectiveness is limited. The evidence on lidocaine as a prokinetic drug is still inconclusive in human medicine.16
The purpose of this study was to analyze the association of pre-, intra-, and postoperative factors with development of POI and to assess the influence of postoperative lidocaine treatment on the incidence of POI and on short-term survival.
Materials and Methods
Studied Data and Definitions
The study population included a total of 157 small intestinal surgical colic cases, presented to the Large Animal Internal Medicine Clinic of Ghent University, Belgium between March 1, 2004 and December 31, 2006. Inclusion criteria were survival for at least 24 hours postoperatively and age ≥1 year; pony breeds were excluded from the study.
Recorded variables included sex, age, and weight of the horse, duration of signs before admission, sedation at home (no sedation, detomidine, romifidine, or an opiate), and sedation at the clinic (no sedation, xylazine, or detomidine). The recorded admission parameters were heart rate, mucous membranes (normal, congested, or cyanotic), capillary refill time (CRT; <2 or ≥2 seconds), skin turgor (normal or decreased), borborygmi on the left side (normal, reduced, or tympanic), borborygmi on the right side (normal, reduced, or tympanic), plasma PCV, blood pH, base excess, presence of reflux (>2 L of reflux upon nasogastric intubation), and quantity of reflux (≤8 or >8 L). The intraoperative variables included preanesthetic sedative (xylazine, detomidine, or romifidine), surgeon (S1 through S6), nature of the lesion (strangulating or nonstrangulating), small intestinal resection, intestinal emptying (defined as the removal of intestinal contents during or after resection, by jejunal or cecal enterotomy), anesthesia duration, lowest mean arterial pressure (MAP), hypotension (defined as MAP < 70 mmHg recorded for at least 15 minutes), dobutamine administration, hypoxemia (Po2 < 95 mmHg [12.65 kPa] in at least 1 arterial blood sample), lowest recorded base excess, and lidocaine administration. The postoperative variables included prophylactic lidocaine treatment and prophylactic metoclopramide treatment for the outcome variable POI; for the outcome variable short-term survival, recorded variables were POI, curative lidocaine treatment, curative metoclopramide treatment, and erythromycin treatment.
Clinical examination at admission encompassed a routine general examination, followed by rectal examination, transabdominal ultrasound,a and passing of a nasogastric tube for performance of gastric decompression. This tube was left in place until the completion of surgery.
During anesthesia, all but 11 horses received a continuous rate infusion (CRI) with lidocaineb (loading dose 1.5 mg/kg followed by 0.033 mg/kg/min). A CRI of dobutaminec was started at a dose of 0.5 μg/kg/min in horses with hypotension and was gradually increased to a maximum dose of 1.97 μg/kg/min, aiming to obtain an MAP > 70 mmHg.
Transabdominal ultrasound to check for gastric distention with fluid17,18 was routinely performed 12 hours after surgery or, according to the clinician's discretion, if a horse had an increased heart rate or was uncomfortable. If gastric liquid content was observed or if the horse remained uncomfortable, nasogastric intubation was performed. In line with earlier reports,3,6 the following case definition of POI was adopted: >20 L of reflux retrieved within 24 hours, or >8 L of reflux retrieved at a single occasion. Short-term survival was defined as survival to discharge from the hospital.
Three different prokinetic drugs were used postoperatively: lidocaine (loading dose 1.3 mg/kg IV followed by a CRI of 0.05 mg/kg/min), metoclopramided (0.05 mg/kg IM q6h) and erythromycine (2 mg/kg IV in 1 L of physiologic solution q12h). Lidocaine and metoclopramide were used as both prophylactic and curative treatments. Erythromycin was only used as a treatment in horses with prolonged POI, when other prokinetics seemed to have failed. Prophylactic prokinetic treatments were started immediately postoperatively, as soon as the horse was stabled, and discontinued if the horse's clinical parameters, blood PCV, and base excess remained within normal limits for 24 hours. For lidocaine, the infusion rate was halved (0.025 mg/kg/min) 24 hours before cessation of CRI.
Descriptive statistics of continuous variables were presented by median and range. For categorical data, frequencies of occurrence were presented.
The relation between the variables listed above and the occurrence of POI was analyzed by logistic regression.f First, the relation of each variable to POI was evaluated in a univariable logistic regression model. Odds ratios (OR), including 95% confidence intervals (CI), were reported for all variables with a P value < .20.
For all continuous variables, the shape of the relationship with the outcome variable was assessed by plotting the log odds of the outcome versus the continuous variable.19 Whenever a nonlinear relationship was observed, the variable was categorized using logical cutoff values based both upon the shape of the plot and biological reasoning.
All variables with a univariable P < .20 were included in a multivariable logistic regression model, which was constructed in a backward stepwise manner. Pearson's and Spearman's ρ correlations between the parameters were determined, and if 2 variables were highly correlated (r2 > 0.60), only the variable with the smallest P value was included in the multivariable model. In the final multivariable model, all 2-way interactions between significant variables were evaluated. Significance was set at P < .05.
Uni- and multivariable models were created in a similar way for short-term survival. The purpose of these models was to evaluate the effect of lidocaine treatment on survival, independent of the presence of POI or other possible confounding factors.
Population Characteristics and Clinical Findings at Admission
The master dataset contained 126 records of horses with small intestinal surgical colic that fulfilled the inclusion criteria. However, certain data could not be retrieved in all horses. The study population included 47 (38%) mares, 45 (36%) geldings, and 32 (26%) stallions. The median age of horses was 9 years (range 1–26 years). The exact breed of the horses was often (n = 64) not recorded; 39 out of 62 horses of known breed were warmbloods. Median body weight was 548 kg (range 250–720 kg).
The duration of signs at admission ranged from 1 to 72 hours, with a median duration of 8 hours. Given the nonlinear relationship among duration of signs at admission and the log odds of POI and short-term survival, it was decided to dichotomize this variable into 2 categories (<20 and ≥20 hours). One hundred and five horses arrived within 20 hours; 21 had a longer duration of signs.
Overall, a total of 40 (32%) horses had been sedated; either by the referring veterinarian (26 out of 96 horses; of which 8 horses with detomidine; 8 with romifidine, and 10 horses with an opiate) and/or at arrival in the clinic (20 out of 123 horses). For the latter category, in the 1st year reviewed, mostly detomidineg (10 μg/kg bwt) was used (n = 12) and later on xylazineh (0.7 mg/kg bwt) was preferred (n = 8). Both at home and at the clinic, primarily the horses with strangulating lesions had been sedated.
The clinical parameters at admission varied widely. Heart rate varied between 36 and 112 beats/min (median 60 beats/min). Mucous membranes were of normal color in 70 horses (57%), congested in 45 (37%), and cyanotic in 8 (6%). The CRT was prolonged in 50/117 horses (43%) and skin turgor was decreased in 47/118 horses (40%). Borborygmi on the left side were normal in only 7 horses (6%), reduced or absent in 107 (87%), and reduced and tympanic upon percussion in 9 (7%) cases. On the right side of the abdomen, tympanic sounds were heard more often (n = 15; 12%). PCV ranged between 25 and 74% (median 40%). Because the relationship between PCV and the log odds of POI and short-term survival was nonlinear and the curve showed a clear bend around 50%, it was decided to dichotomize this variable into 2 categories (<50% and ≥50%). Twenty out of 126 horses (16%) had PCV values ≥ 50%. Blood pH and base excess ranged between 7.21 and 7.55 (median 7.38) and −12 and +11 mEq/L (median +1.9 mEq/L), respectively. Upon nasogastric intubation, 33 horses (26%) had >2 L of reflux; 20 of them were presented with >8 L of reflux.
Before surgery, horses were sedated with xylazine (n = 77; 66%), romifidinei (80 μg/kg bwt) (n = 28; 24%), or detomidine (n = 12; 10%), depending on the anesthetist at service. When sedation had already been performed at the internal medicine department, the same sedative was used as anesthetic premedication.
The study population encompassed 84 (67%) strangulating and 42 (33%) nonstrangulating small intestinal obstructions. A small intestinal resection was carried out in 55 cases (44%); 50 of these horses had strangulating lesions, and 5 were diagnosed with nonstrangulating lesions. An ileal bypass without concurrent resection was performed 5 times. During or after resection, removal of intestinal contents by cecal or jejunal enterotomy was performed in 15/52 (29%) cases, mostly (n = 14) strangulating obstructions. In the remaining 66 horses, only decompression, repositioning and/or manual stripping of small intestinal content into the cecum was necessary, followed by an enterotomy of the cecum in 25 horses.
Duration of anesthesia ranged between 30 and 315 minutes (median 130 minutes). Hypotension (n = 102) and/or hypoxemia (n = 63) was encountered in many cases. Median lowest MAP was 57 mmHg (range 18–105 mmHg) and dobutamine was administered to 98 horses. The lowest recorded intraoperative base excess ranged from −11 to +8 mEq/L (median −2 mEq/L).
Prokinetic drugs were administered postoperatively to 109/124 horses (88%). Lidocaine (n = 77) and metoclopramide (n = 55) treatments were started immediately after surgery for the prevention of ileus (referred to as prophylactic treatment). In 43 of these cases, the combination of lidocaine and metoclopramide was used. Sometimes prokinetic treatment was not given until the horse developed ileus (referred to as curative treatment). In those cases, lidocaine (n = 12) was often combined with other prokinetics. Nine horses received lidocaine and erythromycin, 2 received lidocaine, erythromycin, and metoclopramide and 1 horse was treated with lidocaine and metoclopramide. A few horses received single treatments with metoclopramide (n = 3) or erythromycin (n = 5).
Out of 126 horses that survived longer than 24 hours postoperatively, 41 (33%) developed POI. Frequencies were comparable in the nonstrangulating group (14/42; 33%) and the strangulating group (27/84; 32%). In the group of horses that received prophylactic lidocaine treatment, only 16 out of 77 (21%) developed ileus, while in the untreated group, 24 out of 47 horses (51%) developed POI. In the prophylactic metoclopramide group, 15 out of 55 horses (27%) developed POI; out of 69 untreated horses, 25 (36%) were affected.
The onset for need of gastric decompression ranged from <1 day up to 3 days postoperatively (median 1 day), and horses refluxed for 1 up to 15 days (median 6 days). Nine horses needed a repeat celiotomy, 6 of them due to POI.
Overall, 93/126 horses (74%) survived to discharge; survival rates in the nonstrangulating group and the strangulating group were similar (31/42 horses; 74% and 62/84 horses; 74%, respectively). POI had an adverse influence on outcome: only 14/41 POI-affected horses (34%) survived to discharge.
Univariable and Multivariable Analysis
The variables that passed univariable screening for association with the development of POI are listed in Table 1. PCV, turgor, and presence of reflux were excluded from the multivariable model because of high correlations with heart rate, CRT, and quantity of reflux, respectively. In the multivariable model, the variables heart rate, quantity of reflux (≤8 versus >8 L), small intestinal resection, and prophylactic lidocaine treatment remained significant (Table 2). There were no significant 2-way interactions between these variables.
Table 1. Results of univariable analysis for the identification of risk factors for the development of equine postoperative ileus in 126 surgical small intestinal colic cases (2004–2006).
|Heart rate (bpm)||123||<.001||1.05||1.03–1.08|
|CRT||117||.036|| || |
| <2 seconds (Ref.)||67||—|| || |
| ≥2 seconds||50||.036||2.31||1.06–5.06|
|Skin turgor||118||.11|| || |
| Normal (Ref.)||71||—|| || |
|PCV||126||.002|| || |
| <50% (Ref.)||106||—|| || |
|Presence of reflux (>2 L)||125||.17|| || |
| No (Ref.)||92||—|| || |
|Quantity of reflux||124||.027|| || |
| ≤8 L (Ref.)||104||—|| || |
| >8 L||20||.027||3.02||1.13–8.02|
|Surgeon||119||.16|| || |
| S1 (Ref.)||25||—|| || |
|SI resection||126||.021|| || |
| No (Ref.)||71||—|| || |
|Lowest MAP (mmHg)||120||.053||0.97||0.94–1.00|
|Hypotension||120||.13|| || |
| No (Ref.)||18||—|| || |
|Prophylactic lidocaine||124||.001|| || |
| No (Ref.)||47||—|| || |
Table 2. Results of multivariable analysis for the identification of risk factors for the development of equine postoperative ileus in 120 surgical small intestinal colic cases (2004–2006).
|Heart rate (bpm)||120||.002||1.05||1.02–1.08|
|Quantity of reflux||120||.032|| || |
| ≤8 L (Ref.)||100||—|| || |
| >8 L||20||.032||4.16||1.13–15.4|
|SI resection||120||.001|| || |
| No (Ref.)||67||—|| || |
|Prophylactic lidocaine||120||.013|| || |
| No (Ref.)||46||—|| || |
Concerning short-term survival, the parameters that passed univariable screening are shown in Table 3. In the multivariable model, only POI and lidocaine treatment (prophylactic and curative) remained significant (Table 4). No significant interactions were found between these parameters. Lidocaine-treated horses had 3.33-fold higher odds to survive to discharge (95% CI 1.02–11.1, P= .047) than untreated horses, whereas horses suffering from POI were 28.2 times more likely to die (95% CI 9.11–87.2, P < .001).
Table 3. Results of univariable analysis for the identification of risk factors for nonsurvival in 126 surgical small intestinal colic cases (2004–2006).
|Sex||124||.17|| || |
| Mare (Ref.)||47||—|| || |
|Heart rate (bpm)||123||<.001||1.05||1.02–1.08|
|CRT||117||.017|| || |
| <2 seconds (Ref.)||67||—|| || |
| ≥2 seconds||50||.017||2.81||1.21–6.55|
|Skin turgor||118||.083|| || |
| Normal (Ref.)||71||—|| || |
| <50% (Ref.)||106||—|| || |
| ≥50%||20||.002|| || |
|Presence of reflux (>2 L)||125||.13|| || |
| No (Ref.)||92||—|| || |
|Quantity of reflux||124||.067|| || |
| ≤8 L (Ref.)||104||—|| || |
| >8 L||20||.047||2.73||1.01–7.36|
|Surgeon||119||.078|| || |
| S1 (Ref.)||25||—|| || |
|SI resection||126||.063|| || |
| No (Ref.)||71||—|| || |
|Intestinal emptying||52||.077|| || |
| No (Ref.)||37||—|| || |
|Anesthesia duration (10 minutes)||120||.14||1.05||0.98–1.13|
|Lowest MAP (mmHg)||120||.10||0.98||0.95–1.01|
|Lidocaine in anesthesia||124||.037|| || |
| No (Ref.)||11||—|| || |
|Lidocaine (prophylactic + curative)||124||.074|| || |
| No (Ref.)||35||—|| || |
|Erythromycin (curative)||124||<.001|| || |
| No (Ref.)||108||—|| || |
|Postoperative ileus||126||<.001|| || |
| No (Ref.)||85||—|| || |
Table 4. Results of multivariable analysis for the identification of risk factors for nonsurvival in 124 surgical small intestinal colic cases (2004–2006).
|Lidocaine (prophylactic + curative)||124||.047|| || |
| No (Ref.)||35||—|| || |
|Postoperative ileus||124||<.001|| || |
| No (Ref.)||84||—|| || |
POI in horses in this study was associated with highly increased odds of death. This underlines once again the need for reliable identification of horses at risk and for the development of effective prophylactic and curative treatment protocols for this condition. A high heart rate, the presence of >8 L of reflux at admission, and the performance of a small intestinal resection were identified as risk factors associated with development of POI. Furthermore, the current findings are supportive for a significantly reduced risk of POI in horses that received prophylactic lidocaine treatment. The results also suggest a beneficial effect of lidocaine treatment on survival rates.
Several studies have shown that an impaired cardiovascular state, particularly an increased PCV level, is related to a higher incidence of equine POI.3,4,6,8 This is again confirmed in the present study, in which the variables PCV, CRT, and heart rate all passed univariable screening. Heart rate and PCV were highly correlated (r2= 0.61), and in the final multivariable model only heart rate, the main associated factor, was retained.
The association between the presence of reflux at admission and the occurrence of POI has been demonstrated previously.2 Furthermore, the findings of this study are suggestive for an increased risk of POI in horses presented with >8 L of reflux.
Although use of α2-agonists has been reported to suppress duodenal motility,20 in this study no significant association to POI of sedation or type of sedative used either at home, at admission, or preanesthetically could be demonstrated.
In line with expectations, the performance of a small intestinal resection was strongly associated with an increased risk for development of POI. This highlights, in accordance with earlier findings,1 the importance of timely referral and prompt surgical intervention. In cases where a resection was performed, the removal of intestinal contents via the site of resection or by cecotomy did not have a significant influence on the prevalence of POI. When interpreting these results, one should take into account the limited number of occasions during which this procedure was carried out in the studied population (n = 15). Further research is necessary to gain a clearer view on that matter.
For the medical treatment of POI, 3 types of drugs were used during the study. Erythromycin was only used in very persistent cases and was not significantly associated with outcome in the multivariable model. Metoclopramide had no significant influence on occurrence of POI, in contrast to a previous study.21 However, because of the occasional occurrence of extrapyramidal adverse effects associated with its IV administration, metoclopramide was administered IM and at a much lower dose in this study. This could explain the observed lack of effect.
Interestingly, the positive effect of lidocaine on survival was not related just to its protective effect in preventing POI, for the effect on survival was significant in a multivariable model, independent of POI. Various other properties of lidocaine, such as anti-inflammatory, analgesic, and/or antiendotoxin effects, likely mediated this effect of lidocaine on short-term survival.
Results of studies on lidocaine treatment have heretofore been variable. In vitro, the prokinetic effect of lidocaine was restricted to proximal duodenal smooth muscle strips, with no effect on jejunal or pyloric antral strips.22 No evidence was found for any prokinetic effect of lidocaine treatment on postoperative jejunal motility in healthy horses.23 Prophylactic lidocaine treatment did have some positive effects on intestinal contractility after colic surgery, but there was no effect on the development of gastric reflux or on survival.24 This lack of effect on survival was in accordance with a later study,25 in which POI-affected horses were nevertheless found to exhibit an earlier inhibition of reflux and an earlier discharge after treatment with lidocaine.
The lack of prokinetic effects of lidocaine, encountered in both in vitro studies and in healthy horses, could indicate that the drug exerts an indirect effect on motility by decreasing pain or intestinal inflammation. Lidocaine has been demonstrated to inhibit neutrophil activity26 and attenuate cytokine response, as well as accelerate recovery of bowel function, in humans.27 After colic surgery, significant neutrophilic inflammation occurs in the equine jejunum,28 which supports the hypothesis that intestinal inflammation plays a role in the pathogenesis of POI. The anti-inflammatory properties of lidocaine could mediate the effects on POI and short-term survival that were observed in this study. It has also been demonstrated that systemic lidocaine treatment can protect the equine jejunum against the negative effects of flunixine meglumine,29 which is reported to delay jejunal mucosal recovery after ischemic injury.30 Before induction of anesthesia, as well as postoperatively, all the horses in the present study received flunixine meglumine as an analgesic treatment. It would be interesting to evaluate the effect of lidocaine in the absence of nonselective COX inhibitors.
In 10 out of 47 horses in this study that did not receive prophylactic postoperative lidocaine treatment, lidocaine was not administered during surgery either. Therefore it cannot be ruled out that the demonstrated effects of lidocaine can be partially attributed to intraoperative administration. A trend (P= .084) toward reduced odds of POI has been reported in horses that had been treated intraoperatively with lidocaine.6 More research is necessary for a final conclusion.
Review of the literature reveals substantial variation between studies, with regard to the prevalence and outcome of POI. This can be explained by differences in the studied populations or by different definitions of POI, as well as by variations in treatment protocols. The variables heart rate, quantity of reflux at admission, and small intestinal resection can be a useful aid in identifying horses at risk for POI and in providing a more accurate prognosis. The results of this study are suggesting a protective effect of IV lidocaine treatment against the development of POI and a positive effect on short-term survival. Further research is needed to get more insight into the prokinetic properties and mechanism of action of lidocaine.
aSonos 100, HP, 2.5 MHz sectorial probe, HP Belgium, Diegem, Belgium
bLidocaine (Laocaine), Schering-Plough Animal Health, Segré, France
cDobutamine (Dobutamine EG), Eurogenerics, Brussels, Belgium
dMetoclopramide (Primperan), Sanofi-Synthelabo, Brussels, Belgium
eErythromycin (Erythrocine IV 1 g), Abbot, Louvain-la-Neuve, Belgium
fSPSS 15.0, SPSS Inc, Chicago, IL
gDetomidine (Domosedan), Pfizer Animal Health, Louvain-la-Neuve, Belgium
hXylazine (Xyl-M), VMD, Arendonk, Belgium
iRomifidine (Sedivet), Boehringer Ingelheim, Brussels, Belgium