Ativan may be linked to worse survival outcomes in pancreatic cancer

Providers often prescribe benzodiazepines (BZDs), a class of drugs that can help with anxiety, to people with cancer. Two commonly prescribed BZDs are lorazepam, better known as Ativan, and alprazolam, or Xanax. How these drugs affect survival outcomes in people with cancer has not been extensively studied.

New data analysis published Aug. 17 reveals that people with pancreatic cancer who took lorazepam were 3.83 times more likely to have their cancer recur during the duration of the study than people who didn't take the drug (Clin. Cancer Res. 2023, DOI: 10.1158/1078-0432.CCR-23-0547). In contrast, people who took alprazolam were less likely to have their cancer recur than those who didn’t take alprazolam.

Researchers found this possible link by examining existing data regarding survival outcomes and prescription records from pancreatic cancer patients at Roswell Park Comprehensive Cancer Center who were treated with chemotherapy between 2004 and 2020; 40 that had taken lorazepam, 29 that hadn’t taken lorazepam, 27 that had taken alprazolam, and 42 that hadn’t taken alprazolam.

They then took the research into a mouse models of pancreatic cancer. When they gave lorazepam to the mice, they found that the tumor microenvironment changed in a way that typically makes tumors resistant to chemotherapy.

While historical studies have suggested a possible link between different BZDs and different patient outcomes, “really no one has looked at and tried to understand what effect these might be having on the tumor microenvironment,” says corresponding author Michael Feigin, who studies the molecular causes of cancer at Roswell Park.

The team suspected that a G-protein coupled receptor called GPR68 was at play. GPR68 is commonly expressed on a type of cell in the tumor microenvironment called a fibroblast. When activated, GPR68 spurs greater expression of the cytokine IL-6, which in turn promotes inflammation in the microenvironment and can be tied to tumor progression.

Prior work has suggested that N-unsubstituted BZDs, a class to which lorazepam belongs, are able to activate GPR68, while N-substituted BZDs like alprazolam can’t. Using an assay to test a panel of commonly prescribed BZDs, Feigin’s team was able to confirm this. The difference between the two groups comes down to a nitrogen atom. In N-substituted BZDs, which also include diazepam and temazepam, a particular nitrogen makes three bonds to non-hydrogen atoms. In N-unsubstituted BZDs, like clonazepam, nordiazepam, and oxazepam, the same key nitrogen makes at least one bond to hydrogen.

The ability of lorazepam to flip this switch even though alprazolam can’t makes sense, according to Mary Mader, a medicinal chemist who works at the Indiana Biosciences Research Institute. “The selectivity (or lack of it) can be driven by something as simple as a substitution of an N-H. Alprazolam replaces the N-H and the C=O with an entire triazole ring, which is more than simply substituting N-H with N-CH3,” she says via email.

GPR68 is relevant beyond pancreatic cancer; it’s expressed on fibroblasts in the tumor microenvironment in other cancer types too, Feigin says. His team ended up analyzing data from patients spanning several cancer types. For many of the cancers they found a correlation between lorazepam and worse outcomes.

Lorazepam activating the GPR68 protein could at least partially explain the worse survival outcomes, Feigin says, but more research is needed. He also notes that this is a single center study, and that they need to confirm there’s still a correlation between lorazepam and worse outcomes in data from a larger set of people.

Anirban Maitra, a pathologist at the University of Texas MD Anderson Cancer Center, emphasized that limitation, and noted that people should not change their medications without consulting their oncologists. He also praised the authors’ work, and said larger studies should be done “post-haste.”

“This is a really important study that gives me pause,” he says in an email. “The implications—if validated in larger datasets—would be profound and potentially practice changing.”