SAN DIEGO – However, antibiotics were still prescribed for more than one in three patients tested by any method.
“A positive test by any modality did result in decreased utilization of endoscopy, radiology, and antibiotic prescribing, but this effect appeared to be much greater for the GI PCR assay,” said Jordan Axelrad, MD, speaking at the annual Digestive Disease Week.
“Overall, patients who received GI PCR were 12% less likely to undergo endoscopy, 7% less likely to undergo abdominal radiography, and 11% less likely to be prescribed any antibiotic,” compared with patients who were tested by conventional stool culture, said Dr. Axelrad, a gastroenterologist at New York University.
In a cross-sectional study, Dr. Axelrad and his coauthors looked at patients who underwent stool testing for the 26 months before (n = 5,986) and after (n = 9,402) March 2015, when Dr. Axelrad’s home institution switched from conventional stool culture to the GI PCR panel. For the earlier time period, the investigators included patients who received stool culture both with and without an ova and parasites exam, as well as those who underwent enzyme-linked immunosorbent assay viral testing for rotavirus and adenovirus.
Patient demographic data were included as study variables; additionally, the study tracked utilization of endoscopy, abdominal, or other radiology studies, and ED visits for 30 days after testing. They also included any antibiotic prescribing within the 14 days post testing.
Roughly one-third of patients were tested as outpatients, 1 in 10 in the ED, and the remainder as inpatients. Patient age was a mean 46.7 years for the culture group, and 45.5 years for the GI PCR group.
The multiplex PCR test used in the study tested for 12 gastrointestinal pathogenic bacteria, 4 parasites, and 5 viruses.
As expected, PCR testing yielded a higher positive test rate than conventional stool testing, even when EIA tests were included (29.2% vs. 4.1%). In the 2,746 patients with a positive GI PCR test, a total of 3,804 pathogens were identified. Adenovirus accounted for 39% of these positive results. Positive bacterial results were seen in about 65.0% of the positive subgroup, with Escherichia coli subtypes seen in 51.7% of the positive tests.
Overall, positive results for viruses, bacteria, and multiple pathogens were more likely with GI PCR testing, compared with conventional testing (P = .001 for all). Parasites accounted for only 8.2% of the positive PCR test results, but this was significantly more than the 3.7% seen with conventional testing (P = .011).
At the 14-day mark post testing, “Patients who underwent a GI panel were less likely to be prescribed any antibiotic. But overall, antibiotics were fairly common in both groups,” said Dr. Axelrad, noting that 41% of patients who underwent stool culture received an antibiotic by 14 days, compared with 36% for patients who underwent a GI PCR panel (P = .001).
By the end of 30 days, most patients in each group had not received an endoscopic procedure, with significantly more procedure-free patients in the PCR group (91.6% vs. 90.4%; P = .008).
Against a backdrop of slightly higher overall radiology utilization in the PCR group – potentially attributable to practice trends over time – abdominal radiology was less likely for these patients than for the culture group (11.4% vs. 12.8%; P = .011).
The 30-day ED visit rate was low and similar between groups (11.4% for PCR vs. 12.8% for culture; P = .116).
The much quicker turnaround for the GI PCR panel didn’t translate into a shorter length of stay, though: Inpatient length of stay was a median 5 days in both groups.
“We feel that the outcomes that we noted were likely due to the increased sensitivity and specificity” of the PCR-based testing, said Dr. Axelrad. “Obviously, if you have more pathogen-positive findings, you may be less likely to order extensive testing. And if you’ve identified something like norovirus, you may feel reassured, and not order further testing.”
Dr. Axelrad pointed out that his institution’s overall PCR positivity rates were lower than the 70% rates some other studies have reported. “We feel that, given our large sample size, our results may more accurately reflect clinical practice, and perhaps that lower positivity rate may reflect increased use of this test in an inpatient setting,” he said. “We’re looking at that.”
Study limitations included the retrospective nature of the study. “Also, as we all know, PCR testing fails to discriminate between active infection and asymptomatic colonization,” raising questions about whether a positive PCR test really indicates true infection, noted Dr. Axelrad.
“Coupled with a high-sensitivity rapid turnaround, there’s the potential to reduce costs, but the cost-effectiveness of these assays has not been fully determined. There are several studies looking at this,” with results still to come, he said.
The notable reduction in antibiotic prescribing for those patients who received PCR-based testing means that GI PCR panels could be a useful tool to promote antibiotic stewardship, though Dr. Axelrad also noted that “antibiotics were still used in about a third of all patients.”
Dr. Axelrad reported no outside sources of funding. He has performed consulting services for and received research funding from BioFire, which manufactured the GI PCR assay used in the study, but BioFire did not fund this research.
SOURCE: Axelrad J et al. DDW 2019, Presentation 978.
