Cytomegalovirus in IBD: where to biopsy, whom to treat

Where to biopsy and how many biopsies are needed is the question. “Not every biopsy fragment has inclusions, and not every biopsy from every site has inclusions. Ulcers and granulation tissue have the highest likelihood of having inclusions, so biopsies from ulcerated areas are going to be more productive,” he said. Some studies have shown that up to 16 biopsies may be necessary to demonstrate CMV in a patient who has CMV disease, but it’s possible to reduce that number by focusing on ulcerated areas (McCurdy JD, et al. Inflamm Bowel Dis. 2015;21[12]:2833–2838; Roblin X, et al. Am J Gastroenterol. 2011;106[11]:2001–2008).

McCurdy, et al., found in their retrospective study of 68 patients that biopsy location and number are important when assessing for CMV. In the majority of patients with ulcerative colitis, CMV was detected from the left colon. Among patients with Crohn’s disease, a substantial number had CMV detected from only the right colon.

Some studies have cast doubt on whether IHC is necessary to diagnose viral infections in humans, including CMV in patients with IBD, Dr. Misdraji said. H&E is highly specific (92 to 100 percent) but has variable sensitivity (10 to 87 percent). “We’ve all had that situation where we get a CMV stain and it’s positive,” but the inclusions were not seen on the initial H&E.

IHC improves sensitivity to 78 to 93 percent, he said, though some question its necessity because clinically relevant infections generally have inclusions on H&E (Solomon IH, et al. Am J Clin Pathol. 2017;147[1]:96–104). “But there are always cases in which we just didn’t see them, or they’re atypical inclusions, they’re a little smaller than usual, and you’re surprised by the stain.” The European Crohn’s and Colitis Organization guidelines recommend tissue PCR or IHC for diagnosing CMV in the setting of immunomodulator-refractory IBD.

The question is which patients need to be treated. Here there is growing evidence that CMV disease can be stratified into low- and high-grade disease, Dr. Misdraji said, and that high-grade disease is associated with a higher colectomy rate but may also be more likely to benefit from antiviral therapy. “The big issue for us at this point is that all the studies are using different definitions of what constitutes high-grade disease for us to report,” he said.

Kuwabara, et al., used resection specimens and defined high-grade disease as greater than 10 inclusions in any one histologic section. That was the first study that found an association between high-grade disease and a worse outcome, shorter interval to colectomy, and steroid resistance, Dr. Misdraji said (Kuwabara A, et al. J Gastroenterol. 2007;42[10]:823–829).

A 2011 article on CMV reactivation in ulcerative colitis patients defined high-grade disease as visible inclusions on H&E versus only on the IHC stain (Nguyen M, et al. Ulcers. 2011;2011:282507). A study published in 2015 reported that the beneficial effect of treatment was most notable in patients with high-grade disease, defined as more than five viral inclusions in any one biopsy fragment. “The number 5 was chosen arbitrarily,” the authors wrote, “as a cut-off point based on a subjective assessment of what constituted a heavy viral load” (Jones A, et al. Clin Gastroenterol Hepatol. 2015;13[5]:949–955).

“Once that study showed association, other people looked at the same number,” Dr. Misdraji said, among them Zagórawicz, et al., who found that five or more CMV IHC-positive cells per biopsy section were indicative of a greater colectomy risk, though response to ganciclovir was not clearly better in that group (Zagórowicz E, et al. J Crohns Colitis. 2016;10[10]:1205–1211).

Clos-Parals, et al., in their study of 46 cases, defined high-grade disease as greater than two inclusions by IHC on a single tissue fragment because at five inclusions they found no association with a higher colectomy rate, Dr. Misdraji said (Clos-Parals A, et al. J Crohns Colitis. 2019;13[3]:385–388). Oh, et al., used five inclusions and reported finding no difference in outcome between high- and low-grade CMV disease (Oh SJ, et al. Scand J Gastroenterol. 2019;54[8]:​976–983).

“Looking at the outcome or response to antiviral therapy, the majority of studies show that high-grade disease matters,” he said, “but not all do.” Treatment may be more appropriate for patients with moderate to severe steroid-refractory relapses with a high viral load, he said, “whereas a negative blood CMV PCR test result, a very low quantitative PCR in the tissue, and maybe a single inclusion on IHC can support a decision to abstain from antiviral therapy because there are risks to ganciclovir.”

Ganciclovir is generally given with immunosuppressive therapy, Dr. Misdraji said, “because you want the synergistic effect of reducing the viral replication but also keeping down the inflammation,” especially when anti-TNF-α is used. If CMV disease becomes severe and systemic, immunosuppressive therapy is discontinued.

Dr. Misdraji summed up with the following suggestions for the histologic assessment for CMV:

• Look for CMV inclusions on the H&E stain of severely inflamed or ulcerated biopsies and confirm with IHC. “We’ve all seen ulcers that have funny fibroblasts, and then when you get the stain they’re negative.”
• Order CMV IHC in severely inflamed or ulcerated biopsies, particularly but not exclusively if you are aware that the patient is refractory to their steroid therapy, or if clinically requested.
• Report whether the viral inclusions are visible on the H&E or only on the IHC. “Even if the IHC is positive, I’ll go back to the H&E and report if I can see them on the H&E because that may indicate more severe disease,” he said. “And I will report the number of IHC-positive cells in the single most involved tissue fragment, knowing that a threshold for high-grade disease is not established. I’ll at least give a number or say whether they’re numerous or it’s a single inclusion.”
• Don’t overinterpret lymphocytes or plasma cells that are staining as relevant.

In Fig. 1 is an example of an IHC stain showing numerous inclusions. “They should be large cells with nuclear expression,” Dr. Misdraji said. In Fig. 2 is a plasma cell that shows some cytoplasmic staining. “Most people do carry CMV,” he said, but cytoplasmic staining of small lymphocytes or plasma cells, as in Fig. 2, is not indicative of active CMV disease.

Amy Carpenter Aquino is CAP TODAY senior editor.