Animals
This was a retrospective study of 87 cows that had a main diagnosis of type-4 abomasal ulcer. The cows had been admitted to the Veterinary Teaching Hospital, University of Zurich, from January 1, 1991 to December 31, 2014. The final diagnosis was based on the results of postmortem examination. The results were described in detail [12]. The cows ranged in age from 2 to 10 years (mean ± standard deviation [sd], 4.5 ± 1.5 years). Breeds included Brown Swiss (36), Holstein-Friesian (25), Swiss Fleckvieh (24) and crossbred cattle (2). The duration of illness was < 2 days in 36 cows, 2 to 6 days in 40 cows, 7 to 14 days in 6 cows and > 14 days in 4 cows. The majority of cows (n = 36, 49%) became ill within 4 weeks after calving (Fig. 1); this incidence was significantly higher than that of other stages of lactation (P < 0.01). Twenty-four cows had been treated with non-steroidal anti-inflammatory drugs (NSAIDs), six with corticosteroids and another six with NSAIDs and corticosteroids before referral, but the exact dosages were not known.
Clinical examination
The cows underwent a thorough clinical examination [13]. General health was evaluated by determining demeanour, appearance of hair coat and muzzle, skin elasticity, position of the eyes in the sockets and skin surface temperature. Each cow was observed for signs of pain such as spontaneous grunting and bruxism. The rumen was assessed for degree of fill, number and intensity of contractions and content stratification. Sensitivity in the reticular region was assessed by preventing the animal from breathing for a short period by placing a plastic rectal sleeve over the mouth and nose and listening for grunting during the ensuing deep breath. This was followed by foreign body tests, which included the pole test, back grip and percussion of the abdominal wall over the region of the reticulum with a rubber hammer. Each test was carried out 4 times as described [14], and the reaction of the animal was observed each time. A test was considered positive when it elicited a short grunt at least three out of four times. The response to a test was considered questionable when it elicited a grunt two out of four times and negative when the animal did not grunt or grunted only once. Ballottement and simultaneous auscultation as well as percussion and simultaneous auscultation of the abdomen on both sides and rectal examination were also carried out. Faeces were assessed for colour, consistency, amount, fibre particle length and abnormal contents.
Laboratory analyses
The following blood samples were collected from all cows: 5 ml of EDTA blood for haematological analysis, 10 ml of whole blood for serum biochemistry, 2 ml of whole blood mixed with 0.2 ml heparin for venous blood gas analysis and 5 ml of EDTA blood for the glutaraldehyde test. Haematological analysis included the determination of haematocrit, total leukocyte count and the concentrations of fibrinogen and total protein using an automated blood analyser (CELL-Dyn 3500, Abbott Diagnostics Division, Baar). The concentration of serum urea nitrogen was determined at 37 °C using an automated analyser (Cobas-Integra-800-Analyser, Roche Diagnostics, Basel) and the manufacturer’s reagents (Roche Reagents) according to the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC). Venous blood gas analysis was done using an automated analyser (RapidLab 248, Siemens Schweiz AG, Zurich). A glutaraldehyde test (Glutaltest®, Graeub AG, Bern) was done according to the manufacturer’s instructions. Results were interpreted relative to reference intervals recently reported [15].
A urine sample, mainly collected during spontaneous micturition, was analysed in 76 cows. The colour and transparency of the urine were assessed macroscopically, and the specific gravity was determined using a refractometer (HRMT 18, A. Krüss Optronic GmbH, Hamburg, Germany). A urine test strip (Combur9®, Roche, Basel) was used to determine pH and the presence of protein, erythrocytes, glucose, ketones, leukocytes, nitrite, urobilinogen and bilirubin.
A sample of rumen fluid (200 to 300 ml) was collected using a Dirksen probe [13] in 67 cows and assessed for colour, odour, consistency and pH. In addition, the methylene blue reduction time and the concentration of chloride were determined. The concentration of chloride in rumen fluid was carried out using an MK-II-Chloride Analyser 9265 (Sherwood, Cambridge).
Ultrasonographic examination and abdominocentesis
Seventy-five cows underwent ultrasonographic examination of the reticulum (n = 58), abomasum (n = 21) and abdomen (n = 63) using a 3.5 or 5.0 MHz convex or linear transducer [16]. Echogenic deposits, with or without hypoechoic or anechoic fluid inclusions, and structures of various shapes and echogenicities with central echogenic fluid collections, reflect inflammatory changes of the peritoneum that include fibrinous deposits and abscesses were interpreted and referred to as inflammatory lesions of the peritoneum [16, 17].
Amount and appearance of fluid accumulations and fibrin strands were noted. Forty-seven cows with ultrasonographic evidence of abdominal fluid accumulation underwent ultrasound-guided abdominocentesis [18] and the aspirated fluid was assessed macroscopically with respect to colour, odour and transparency. Specific gravity and total solids were measured with a refractometer. The aspirated fluid was considered an exudate when at least one of the following criteria was met: specific gravity > 1.015, total solids > 30 g/l, cloudy appearance, malodourous appearance and green discolouration.
Euthanasia
All severely affected cows that did not die spontaneously were euthanased immediately after the initial examination or after 2 to 4 days of unsuccessful treatment. Euthanasia was done with pentobarbital (Esconarkon, Streuli Pharma), 80 mg/kg body weight intravenously.
Statistical analysis
The program IBM SPSS Statistics 22.0 was used for analysis. Frequencies were determined for each variable. The Wilk-Shapiro test was used to test the data for normality. Means ± standard deviations were calculated for normal data (total protein) and medians for non-normal data (heart rate, respiratory rate, rectal temperature, haematocrit, white blood cell count, fibrinogen and urea concentration, glutaraldehyde test coagulation time, pH, pCO2, HCO3− and base excess of venous blood, urine pH, urine specific gravity). Differences in ulcer occurrence at various stages of lactation were analysed using a one-way analysis of variance and the post hoc Bonferroni test. A value of P < 0.05 was considered significant.