The medical and surgical indexes at our institution were cross-referenced in order to identify all patients who had been diagnosed with Ogilvie syndrome following a hip arthroplasty in the ten-year period between 1984 and 1993. The diagnosis of Ogilvie syndrome was based on clinical and radiographic evidence of distention of the large bowel in the absence of demonstrable mechanical obstruction. The specific study period was chosen in order to allow a review of the outcomes of this complication during the era of modern colonoscopic decompression. This therapeutic modality had become established as an effective treatment option by the early 1980s10. During the selected period, 10,468 hip arthroplasties were performed. We identified thirty patients (twenty-three men and seven women) in whom acute colonic pseudo-obstruction occurred in the immediate postoperative period. The mean age at the time of the operation was 74.3 years (range, fifty-six to ninety years) for the total group, 74.3 years (range, fifty-six to ninety years) for the men, and 74.0 years (range, sixty-seven to eighty-five years) for the women. Of the 10,468 hip arthroplasties, 7351 were primary procedures. The primary procedures included eighty-seven hemiarthroplasties with cement, 1005 hemiarthroplasties without cement, 3249 total hip arthroplasties with cement, 1949 total hip arthroplasties without cement, 992 so-called hybrid total hip arthroplasties (insertion of the cup without cement and insertion of the stem with cement), sixty so-called reverse hybrid total hip arthroplasties (insertion of the cup with cement and insertion of the stem without cement), and nine primary procedures of unknown type. The remaining 3117 procedures were revisions. Six hundred and ninety-four operations were performed as a result of a traumatic injury, 9766 were elective procedures, and the reasons for eight were unknown. The 10,468 procedures were performed in 5625 women and 4843 men.
The clinical notes and radiographs for each patient in the study group were reviewed. Information that was gathered from the clinical records included demographic details, type of hip arthroplasty performed, type of anesthesia used, previous abdominal operations, underlying medical conditions, types of medications used in the preoperative and postoperative periods, electrolyte values at the onset of the disease, time of onset of symptoms of Ogilvie syndrome, type of symptoms at presentation, date of diagnosis, physical findings at the time of diagnosis, date that treatment was initiated, type of treatments used, duration of symptoms, date of resolution of symptoms, date of return to a general diet, associated complications, and duration of postoperative hospitalization. Information obtained from review of the radiographs included the number of abdominal radiographs made, maximum amount of dilatation of the large bowel, site of maximum dilatation, and date on which maximum dilatation occurred. Maximum colonic dilatation was determined by measuring the maximum width of the colon on anteroposterior abdominal radiographs.
Statistical analysis was performed to identify any associations between the treatment modalities used and resolution of the colonic pseudo-obstruction. The exact Wilcoxon test for categorical data was used to test for associations between ordered categorical variables. The associations between continuous variables were estimated with use of Pearson correlation coefficients.
The prevalence of acute colonic pseudo-obstruction following hip arthroplasty at our institution during the ten-year period of study was 0.29 per cent (thirty of 10,468). Among the thirty patients, men outnumbered women approximately three to one (twenty-three men and seven women); thus, women were at a significantly lower relative risk for the development of acute colonic pseudo-obstruction (0.26 with a 95 per cent confidence interval of 0.13 to 0.61). There was no significant difference between the mean ages of the men and women (p = 0.92). The hip arthroplasty was elective in twenty-two patients and was performed following a traumatic event in eight. The patients who had had an elective procedure were at a significantly lower relative risk for acute colonic pseudo-obstruction than those who had had an operation following a traumatic injury (0.20 with a 95 per cent confidence interval of 0.08 to 0.44). The patients who had had a traumatic injury were managed with bed rest for a mean of seventy hours (range, fourteen to 240 hours) before the operation; those patients who had had an elective procedure were allowed to walk until the time of the operation.
The index procedures included twenty-two primary hip arthroplasties: eight total hip arthroplasties with cement, three total hip arthroplasties without cement, three so-called hybrid total hip arthroplasties, seven hemiarthroplasties without cement, and one hemiarthroplasty with cement. The remaining eight index procedures were revisions: three total hip arthroplasties without cement, one hybrid total hip arthroplasty, two replacements of the femoral component without cement, one replacement of the femoral component with cement, and one exchange to a bipolar head (Fig. 1). No defects of the medial wall of the acetabulum were noted during any revision procedure. With the numbers available, we could detect no significant difference in the relative risk of the development of acute colonic pseudo-obstruction between patients who had had a primary operation and those who had had a revision operation (1.2 with a 95 per cent confidence interval of 0.5 to 2.6). However, among patients who had had a primary procedure, those who had had a hemiarthroplasty were at increased relative risk compared with those who had had a total hip arthroplasty (3.2 with a 95 per cent confidence interval of 1.4 to 7.8). All of the primary total hip arthroplasties were performed on an elective basis, but six of the eight primary hemiarthroplasties were performed following acute trauma and another was done because of a non-union of a fracture of the femoral neck. We could detect no significant differences in the relative risks between the patients who had had a primary arthroplasty with cement and those who had had one without cement (1.60 with a 95 per cent confidence interval of 0.43 to 6.02), between those who had had a primary arthroplasty with cement and those who had had a primary hybrid arthroplasty (0.81 with a 95 per cent confidence interval of 0.22 to 3.06), or between those who had had a primary arthroplasty without cement and those who had had a primary hybrid arthroplasty (1.97 with a 95 per cent confidence interval of 0.40 to 9.72).
General anesthesia was used in eighteen patients; spinal anesthesia, in nine; and epidural anesthesia, in one. Two patients were given spinal anesthesia initially but were given general anesthesia during the operative procedure.
Underlying medical conditions in the thirty patients included diabetes (two patients), hypothyroidism (three patients), and diseases of the central nervous system (nine patients).
Presentation
The onset of symptoms occurred a mean of three days (range, one to eleven days) postoperatively, with the initial symptoms developing on the second, third, or fourth postoperative day in 70 per cent (twenty-one) of the patients. The most common presenting symptom was abdominal distention, which occurred a mean of 3.5 days (range, one to eleven days) postoperatively and was noted in twenty-seven of the thirty patients. Other presenting signs and symptoms included nausea (fourteen patients), vomiting (eight), and abdominal pain (two).
The diagnosis of acute colonic pseudo-obstruction was made a mean of one day (range, zero to three days) after the onset of symptoms. While only twenty-seven patients were noted to have abdominal distention at the onset of symptoms, by the time of the diagnosis of acute colonic pseudo-obstruction all thirty patients had abdominal distention. However, only six had abdominal tenderness. Nineteen of the twenty-five patients for whom information was recorded had had flatus between the time of the arthroplasty and the time of the diagnosis, and twenty-one of the twenty-six for whom the information was recorded had had at least one bowel movement. Consultations with the gastroenterology or general surgery services were requested for all patients, a mean of 0.9 day (range, zero to three days) after the onset of symptoms. Nine patients were evaluated by both consulting services. At the time of diagnosis, eighteen patients had abnormal electrolyte levels: seven patients had hyponatremia, seven had hypokalemia, and four had both hyponatremia and hypokalemia.
A review of radiographs demonstrated that maximum colonic dilatation (mean, 12.2 centimeters; range, 6.5 to 16.0 centimeters) occurred an average of 1.8 days (range, zero to nine days) after the onset of symptoms. Maximum dilatation occurred in the cecum in eighteen patients, in the transverse colon in nine patients, and in the descending colon in one patient.
Treatment
The medical modalities used to treat the acute colonic pseudo-obstruction included the cessation of oral intake, the use of a nasogastric tube, the administration of peristaltic stimulants, and the discontinuation of administration of narcotic medications in most patients (Table I).
After failure of non-operative treatment, colonoscopy was performed in twenty-five of the thirty patients a mean of two days (range, zero to five days) after the onset of symptoms. Five patients had two colonoscopies, and one patient had three colonoscopic decompressions. Three patients had cecostomy because of persistent, massive dilatation despite colonoscopic decompression. No perforations occurred in the thirty patients.
Associated Medical Complications
In addition to the acute colonic pseudo-obstruction, twenty-one other medical complications occurred in fifteen of the thirty patients in the postoperative period. Eighteen of the twenty-one medical complications developed after the onset of the acute colonic pseudo-obstruction. Deep infection developed in three patients, one of whom had positive intraoperative cultures at the time of the index revision operation and probably had a persistent, rather than a new, infection. The deep wound infection was caused by coagulase-negative Staphylococcus in two patients, and the third patient had a polymicrobial infection with coagulase-negative Staphylococcus, Enterococcus, and Corynebacterium.
Four patients died during hospitalization. However, none of these deaths were due to perforation of the bowel. Three deaths resulted from a pulmonary embolism, and the fourth occurred after a cerebrovascular accident. A pulmonary embolism also developed in a fourth patient, who survived. Of the four patients who had a pulmonary embolism, two had it in the immediate postoperative period (one and three days after the arthroplasty). Both patients had received routine postoperative anticoagulation treatment with warfarin, but a therapeutic level of anticoagulation had not yet been attained. The acute colonic pseudo-obstruction was noted after the development of the pulmonary embolism in one of these patients, whereas in the other the pulmonary embolism developed after the onset of the pseudo-obstruction. In the third patient, the pulmonary embolism developed on the ninth postoperative day, five days after the colonic pseudo-obstruction had developed. The patient had been managed with aspirin for anticoagulation, but the administration had been discontinued when oral intake was stopped. In the fourth patient, the pulmonary embolism developed on the fifteenth postoperative day, six days after the onset of the colonic pseudo-obstruction. This patient had been managed with warfarin (first orally and then through a nasogastric tube) throughout the postoperative period, but therapeutic levels of anticoagulation had not been maintained.
Upper gastrointestinal bleeding occurred in three patients. Each of these patients had had postoperative anticoagulation with warfarin, but at the time of the bleeding episode none had prothrombin times beyond the range routinely attained for prophylaxis against deep-vein thrombosis after hip arthroplasty at our institution. Two of the three patients had a major bleeding episode that resulted in changes in vital signs. Both episodes occurred after the onset of the acute colonic pseudo-obstruction, and endoscopic examination revealed that a peptic ulcer (one duodenal and one gastric) had developed in both patients. At the onset of the bleeding problems, the prothrombin times in these patients were 12.6 and 18.0 seconds (normal range, 8.4 to 12.0 seconds). However, both patients were receiving non-steroidal anti-inflammatory medications simultaneously. One patient had received Toradol (ketorolac tromethamine) intramuscularly for control of pain every six hours over a forty-eight-hour period, and the other patient had taken enteric-coated aspirin twice daily. The third upper gastrointestinal bleeding episode occurred in a patient in whom gastritis developed on the first postoperative day. However, this patient was hemodynamically stable and had no additional bleeding episodes. The patient had received the routine dose of warfarin on the night of the operation, but the prothrombin time never increased. The gastritis preceded the onset of the acute colonic pseudo-obstruction.
The other associated medical complications included urinary retention (three patients); urinary tract infection (two patients); and angina, delirium, congestive heart failure, atrial fibrillation, ascites, and thrombocytopenia (one patient each).
Resolution
In the twenty-six patients who survived, the abdominal distention resolved a mean 9.7 days after the onset of symptoms. These patients returned to a general diet a mean of 11.2 days after the onset of symptoms. The fifteen patients in whom other medical complications did not develop stayed in the hospital for a mean of only thirteen days (range, six to twenty-nine days) after the onset of symptoms and a mean of four days (range, two to six days) after the resolution of the distention. In contrast, the eleven patients who had other medical complications and survived stayed in the hospital for a mean of thirty days (range, eleven to fifty-two days) after the onset of symptoms and a mean of seventeen days (range, one to forty-four days) after the resolution of the distention. With the numbers available, we were unable to identify an association between the length of time from the onset of symptoms until colonoscopy and either the number of days until the resolution of symptoms (p = 0.22) or the number of days until discharge from the hospital (p = 0.26). We were also unable to identify an association between the anatomical location (rectum and sigmoid colon, descending and transverse colon, or ascending colon and cecum) at which the colonoscopy was performed and either the number of days before the resolution of symptoms (p = 0.33) or the number of days before discharge from the hospital (p = 0.29).
During hospitalization, the thirty patients had an average of seventeen abdominal radiographs (range, three to fifty radiographs) before resolution of the distention or before death.
Acute colonic pseudo-obstruction has been reported to occur in association with many medical conditions and operative procedures. To our knowledge, the reports in the orthopaedic literature have described only small groups of two to seven patients, in whom the condition developed after a variety of procedures including elective total hip arthroplasty, laminectomy, and internal fixation of fracture of the hip5,15,16. The limited size of these series precluded statistical analysis of risk factors associated with the complication, and information regarding the prevalence of the complication was not presented. In the current study, the low prevalence of Ogilvie syndrome (0.29 per cent) following total hip arthroplasty probably is an underestimation of the true prevalence. It is likely that the clinical presentation of acute colonic pseudo-obstruction with massive colonic dilatation represents one end of a continuum that begins with a milder ileus. Less severe presentations may have been handled solely by the orthopaedic services and thus were not identified by our screening system.
Previous studies have suggested that elderly, debilitated patients who have conditions such as hypothyroidism and diabetes are at increased risk for the development of this complication3,14. The patients in the current series were, on the average, elderly (mean age, 74.3 years), and no patient who was less than fifty-five years old had the complication. However, we did not identify a high rate of diabetes (two of thirty patients) or hypothyroidism (three of thirty patients). The current study documents a lower relative risk of the development of Ogilvie syndrome in female patients and in patients who had an elective procedure. The higher risk in men has been reported previously by Jetmore et al. and by Wanebo et al.: men outnumbered women thirty-three to fifteen and eighteen to five, respectively. We also documented an increased relative risk in patients who had had a primary hemiarthroplasty in comparison with those who had had a primary total hip arthroplasty. This is probably due to the previously noted lower relative risk that was associated with elective procedures, since most of the hemiarthroplasties in this series were performed following acute trauma. Also, the patients who had the operation following trauma had been managed with bed rest for a mean of seventy hours (range, fourteen to 240 hours) preoperatively rather than being permitted to walk until the time of the operative procedure. It is possible that prolonged immobilization contributed to the increased risk of acute colonic pseudo-obstruction in these patients. We did not find any difference in the relative risk between patients who had a primary operation and those who had a revision operation or among patients who had a primary total hip arthroplasty with cement, without cement, or with a so-called hybrid technique.
Massive abdominal dilatation is the cardinal feature of acute colonic pseudo-obstruction. In the current series, twenty-seven of thirty patients had abdominal distention. Nausea and vomiting were less prevalent (fourteen and eight patients, respectively). Symptoms typically developed two, three, or four days after the operation. Before the diagnosis, twenty-one patients had had bowel movements or documented flatus. Similar observations have been noted in other series3,6,8,11,14,17. Non-operative management of acute colonic pseudo-obstruction has been described well previously1,3,9,11,13. It has been suggested that, if these measures fail to alleviate the colonic distention, decompression by means of colonoscopy or cecostomy should be done to prevent perforation when the cecal diameter exceeds nine to twelve centimeters1,3,14,16-18.
The rate of associated medical complications was high in the present series. Twenty-one medical complications other than colonic pseudo-obstruction developed in fifteen of the thirty patients during hospitalization. It is important to note that eighteen complications occurred after the onset of the acute colonic pseudo-obstruction. The related complications included three fatal pulmonary embolisms and one fatal cerebrovascular accident. (A fourth, non-fatal pulmonary embolism also occurred.) Acute colonic pseudo-obstruction, the treatment of which included cessation of oral intake, may have contributed directly to suboptimum oral anticoagulation in two patients who died. Of note, the current study period predated the availability of injectable low-molecular-weight heparin, which is now an effective option for postoperative anticoagulation therapy when oral intake has been restricted2,7. Deep infection was another serious complication in our series. Early infection developed at the site of the arthroplasty in two (7 per cent) of the twenty-nine patients who did not have evidence of infection intraoperatively. Although neither infection was caused by organisms that are most likely to be introduced by insertion of an instrument into the gastrointestinal tract, it is possible that acute colonic pseudo-obstruction indirectly caused these infections. Prolonged use of intravenous access sites for hydration may have provided a portal of entry for skin flora.
In the current study, 13 per cent (four) of the thirty patients died, a high mortality rate. No death resulted from perforation of the bowel; rather, all were due to other medical complications. This mortality rate is similar to the 17 per cent rate (eight of forty-eight) reported by Jetmore et al. for a series in which colonoscopic decompression had been used. It is lower than the 31 per cent rate (thirty-nine of 124) noted by Soreide et al. in a review of the literature that preceded the era of colonoscopic decompression. No perforations occurred in the current study; however, a rate as high as 15 per cent (fifty-two of 351) has been reported9. Wojtalik et al. reported that seven of sixteen patients who had a perforation died. Prompt diagnosis and colonoscopic decompression and cecostomy as needed may have prevented perforation in the current series.
The patients in the present study were hospitalized for an extended period (mean, twenty-one days; range, eight to fifty-five days). Much of the hospitalization was due to the development of other medical complications. The fifteen patients who did not have other medical complications were discharged from the hospital a mean of four days (range, two to six days) after resolution of the acute colonic pseudo-obstruction. The eleven patients who had other complications and survived were discharged at a mean of seventeen days (range, one to forty-four days). Prompt recognition by the orthopaedic surgeon of acute colonic pseudo-obstruction, with early consultation with and intervention by a gastroenterologist or a general surgeon, may reduce both the costs associated with prolonged hospitalization and the prevalence of other medical problems during immobilization. It is therefore important for the orthopaedic surgeon to be aware of this condition, to recognize the characteristic presenting feature following hip arthroplasty, and to understand the potential for notable morbidity and mortality.