In emergency veterinary practice, gastrointestinal foreign body (GFB) removal is a common procedure that is performed with different techniques, such as endoscopy or surgery. The aims of this retrospective, multicentre, clinical study were to report the common locations and types of objects recovered and to investigate clinical factors and outcomes in dogs after surgical or endoscopic treatment for GFB removal. Records of dogs with a GFB diagnosis referred to the Teaching Veterinary Hospital or treated in three different veterinary hospitals from September 2017 to September 2019 were examined. The data obtained from each case included breed, age, clinical signs at presentation, duration of clinical signs, type and location of the GFB, treatment, length of hospitalisation and outcome. Seventy-two dogs were enrolled in the study. There were 42 males (58%) and 30 females (42%). The median age was 36 months (range: 3 months to 8 years). Endoscopic retrieval was performed in 56% of GFBs (located in the stomach or duodenum), whereas 44% of dogs underwent surgery. The type of FB detected varied greatly: kid toy (14%), metallic object/coin (13%), cloth (13%), sock (8%), ball (8%), plastic material (8%), peach stone (7%), fishhook (6%), sewing needle (4%), hair tie (4%), pacifier (3%), plant materials (3%) and others (9%). Moreover, the FBs were classified as sharp (13%, n = 9), pointed (33%, n = 24), blunt (26%, n = 19), or linear (28%, n = 20). In this study, 68% of FBs were localised in the stomach, 25% in the intestinal tract (50% duodenum, 28% jejunum, and 22% ileum), and 7% in both the stomach and small intestine. The type of GFB was not significantly associated with age, site or breed. There was a significant association between the type of GFB and sex: if the dog was male, there was a 38% probability of ingesting linear GFBs. The dog survival rate was 100% in cases treated by gastric endoscopic or surgical removal, 94% in cases treated with enterotomy and 33% in cases in which enterectomy was necessary. Enterectomy and multiple surgical sites were associated with a poor outcome. The presence of vomiting for more than 24 h was significantly associated with death.