TTHealthWatch is a weekly podcast from Texas Tech. In it, Elizabeth Tracey, director of electronic media for Johns Hopkins Medicine in Baltimore, and Rick Lange, MD, president of Texas Tech Health El Paso, look at the top medical stories of the week.
This week’s topics include pregnancy outcomes after kidney donation, an artificial intelligence (AI)-integrated model to manage epilepsy, targeting KRAS for a pancreatic cancer vaccine, and glucose monitoring in non-insulin users with type 2 diabetes.
Program notes:
0:35 Continuous glucose monitoring in type 2 diabetes patients not using insulin
1:40 Decreased HbA1c
2:35 Already shown to be beneficial in those using insulin
3:30 Pancreatic cancer vaccine
4:30 Disease-free survival endpoint
5:32 If T-cell population increased
6:30 Pregnancy outcomes in women who donate kidneys
7:30 Matched 4:1 with nonkidney donors
8:30 Timing of kidney donation
9:15 AI collaboration for managing epilepsy
10:15 Links electroencephalography (EEG) to AI and smartphone app
11:15 Patients requested continuation
12:45 End
Transcript:
Elizabeth: Targeting KRAS in different types of cancer.
Rick: Pregnancy outcomes in women that have donated a kidney.
Elizabeth: Utilization of AI and EEG in taking care of people with uncontrolled epilepsy.
Rick: And is there a role for continuous glucose monitoring in people with type 2 diabetes?
Elizabeth: That’s what we’re talking about this week on TTHealthWatch, your weekly look at the medical headlines from Texas Tech University Health Sciences Center in El Paso. I’m Elizabeth Tracey, a Baltimore-based medical journalist.
Rick: And I’m Rick Lange, president of Texas Tech Health El Paso.
Elizabeth: Rick, I’m going to kick the ball to you, to the New England Journal of Medicine, and take a look at this issue of continuous glucose monitoring in folks with type 2 diabetes who are not using insulin.
Rick: This particular study is for people that are just on oral medications. Or it could be an injectable, but they’re not taking insulin. Because individuals with either type 1 or type 2 diabetes that are on insulin, glucose monitoring continuously has been proven to be helpful. Its role in the treatment of people with type 2 diabetes not on insulin required further study.
And that’s what this study attempted to do. It’s a 26-week trial in 22 primary care practices in the United States. Individuals were randomized to just routine care or to continuous glucose monitoring. There were 145 participants that received continuous glucose monitoring and 138 who just had routine care.
Those that had continuous glucose monitoring, their hemoglobin A1c decreased from 8.8 to 7.2. With routine care, 8.7 to 8.0. They looked at, OK, how often were you in the normal range? They were 21 absolute percentage points more likely to be in the normal range if they were receiving continuous glucose monitoring than if they didn’t. Were they at an increased risk of developing hypoglycemia? And the answer is they were not.
Elizabeth: A very strong argument, of course, for the employment of this technique. And I would note that we are seeing devices that are way easier to use and integrations with watches and cell phones and the provision of that data back to the practice. And so, to me, this makes a lot of sense. What was the downside?
Rick: Really no downside at all. What they weren’t able to ascertain is how did they get their hemoglobin A1c lower. Was it a change in medications? Or were people just more fastidious about either taking their medications or their diet? Or did they change their exercise? But downside? None. And as you mentioned, it’s routinely available, and it’s been already shown to be beneficial in people on insulin. I think we ought to be extending this to all of our patients that have type 2 diabetes, regardless of whether they’re on insulin or not.
Elizabeth: And I would also note, of course, that people become really accustomed to monitoring their own numbers. And I’m wondering if, at some point — and I’m not sure we’ve seen this demonstrated yet in any study — if there’s going to be fatigue after a while at always scrutinizing that data and then making choices based on it.
Rick: Yeah. I can’t answer that for an extended period of time. The individuals in this 6-month study, the median use was 97% of the time. And 86% of the participants used continuous glucose monitoring at least 80% of the time, so they stuck with it at least during this 6-month period.
Elizabeth: This may argue for implantable monitors.
On to then Nature Medicine and something the [European Society for Medical Oncology] meeting just concluded, something much in the news, of course, was the use of vaccines for melanoma and this huge reduction, over 40% reduction in recurrence among the folks in whom that was used. This study is looking at lymph node-targeted KRAS-specific vaccines in pancreas cancer, another bad actor.
So what they did in this study was look at an agent that’s called camupepimut. And this, of course, brings the question to my mind of how in the world do people create these names for these? I wish they’d reveal that, honestly, in these studies, because this is certainly a tongue twister. And this is an agent that drives mutated KRAS-specific T-cell responses.
They were looking at people who had pancreatic adenocarcinoma or pancreas cancer, and they randomized them 2:1 to receive this agent or standard-of-care observation. In their intervention arm, they had 96 people and they were looking at disease-free survival as their endpoint.
After 6 months, disappointingly, did not meet their disease-free survival endpoint in the folks who received it. However, they did see that there were some benefits that corresponded to the use of this agent. Those included this baseline circulating tumor DNA positivity. So, of course, we’re looking for that in people who have cancer, increasingly looking for it as clinically informative, I would note. In this case, it was beneficial with regard to that. And they also were able to show that their KRAS-specific T-cell responses were robust. They were 9.17-fold over baseline. They look at this as efficacy and immunogenicity signals that will help them to advance this therapy in people who have pancreas cancer. And they’re looking at a confirmatory phase III trial.
Rick: There was a robust response in some patients, but not all. But there seemed to be a dose-survivor relationship in that if the T-cell population did not increase ninefold, they didn’t see an improvement in disease-free survival. But in those individuals in whom there did seem to be a response, it was because the T-cell population increased by at least 9-fold and some of them as much as 50-fold.
Elizabeth: By the authors’ admission, they do say that the patients who had cell-free DNA positivity were overrepresented in the treatment arm. And that is kind of an imbalance that may have skewed some of the data, so they really need to take a look at that in groups that are balanced.
Rick: Absolutely. The other thing is, in pancreatic cancer, a significant number of individuals, 75-80%, have KRAS mutations. These mutations are frequently noted in solid cancers as well. About 35-45% of colorectal cancers and about 25% of non-small cell lung cancer.
Elizabeth: Exactly. Moving on then to JAMA.
Rick: Pregnancy outcomes in women that have donated kidneys. If a patient has kidney failure and they’re awaiting kidney transplantation, the transplant lowers their long-term mortality by 55% compared to being on dialysis.
In 2024, there were about 41,000 living kidney donations reported across 88 countries. And oftentimes, those may be family members or nonfamily members. Sixty-five percent of the living kidney donors are female and 75% are of childbearing age. So when a woman donates her kidney, does that change her risk of pregnancy afterwards?
This Canadian study — and by the way, a shout-out to our long-time listener, Tom Jackson, who’s Canadian — included female kidney donors aged 18-42 who had donated a kidney between July of 1992 and August of 2021. And they subsequently had pregnancy of at least 20 weeks gestation. And they matched them 4:1 with women who were matched on their delivery date, their age of delivery, their years of when they delivered. Four of those to every one kidney donor. How often were the women hospitalized with a diagnosis of gestational hypertension, preeclampsia, or eclampsia? Or they looked at 12 other maternal-fetal adverse outcomes.
In the 385 pregnancies in 252 women who had donated a kidney, the risk of being hospitalized for gestational hypertension, preeclampsia, or eclampsia was about 12% versus 8% in women who hadn’t donated a kidney. It appeared that they were more likely to have a small-for-gestational-age child, about 14% versus 8%. And it also appears to be more common with the first pregnancy than it does with the second pregnancy. Counseling donor candidates about this, who may become pregnant, is really important.
Elizabeth: This is a tough balance, isn’t it? Because clearly donating a kidney is going to be a lot easier to deal with the younger you are. The idea that maybe women ought to bear the children they want to bear and then consider kidney donation. And, of course, it’s really hard to also make that determination. Because if you have a relative who needs a kidney, they don’t do it when it’s convenient.
Rick: Yeah. You can look at it either way. You can say there’s an increased risk, and that’s true. Or you could say, well, there’s 86% of those women were not hospitalized for gestational hypertension, eclampsia, or preeclampsia. And there are ways to treat them. Of all the living related donors in this particular study, two-thirds of them were first-degree relatives. It isn’t trying to persuade someone to give or not give a kidney, but just be aware.
Elizabeth: I will be interested in more long-term data relative to this on potential negative outcomes. And we have talked about studies in the past for everybody who donates a kidney, not just this particular cohort.
Back to Nature Medicine, the study of human-artificial intelligence collaboration for comanaging epilepsy in real time. I’m surprised nobody has tried this before. It’s the implementation of a number of different technologies in order to try to deal with this problem, which is frequently quite intractable.
The authors remind us that one-third of the world’s 70 million people with epilepsy have seizures that are uncontrolled by medication. Frequently, of course, implantable devices are considered. They unfortunately are not able, these devices that we have currently, to share this information with patients in real time. And the idea is that this is a missed opportunity because it doesn’t allow the devices to learn from clinical events and teach patients about how their behaviors may affect their health.
So they developed a device platform that converses with patients to help them manage their own epilepsy. And in these folks, who were all inpatients, this prototype links scalp and intracranial EEGs [electroencephalographies] to a secure LLM [large language model]-based artificial intelligence model that communicates directly with their smartphone app and with the patient. They process this model — biomarkers of sleep, medication, cognition, seizure probability — and notifies patients of clinical events and trends, records their symptoms and responses, and uses their feedback to retrain the seizure detector. A very small number of folks in this study. Those who had the scalp EEG, 11. Intracranial monitors, two of them. And as I noted, they were all inpatients during this.
They actually were able to demonstrate that this model could potentially be helpful. They were able to train it so that it really radically reduced its false positives. A lot of the patients at the end of the study said, hey, I’d like to continue this particular thing. My biggest question about it is, yes, and… . If it alerts you that you’re about to have a seizure, what’s the intervention?
Rick: Yeah. On the one hand, it’s providing you information about how you’re living. On the other hand, you’re providing information about what may precipitate a seizure. With that, it begins to learn more about you and provides the information. And you’re right. It doesn’t create an intervention. It just tries to inform you about how to prevent it.
It’s one of the first bidirectional machines. For example, let me say you didn’t sleep very well, and you may figure out that sleep deprivation is what triggers you. It could be that flashing lights trigger you. It can inform you so you can prevent that.
Elizabeth: Right. And the authors say prompt action with timely alerts from the system, like taking additional rescue medication. I haven’t really heard that that’s particularly helpful in staving off a seizure that’s about to take place.
Rick: Yeah. And it may not be for most individuals, but there could be an occasional individual in whom that’s exactly what you do. But you’re right. You have all this information, and unless it’s going to lower your seizure burden, is it useful or not? That’s the next step.
Elizabeth: Time will tell. On that note, that’s a look at this week’s medical headlines from Texas Tech. I’m Elizabeth Tracey.
Rick: And I’m Rick Lange. Y’all listen up and make healthy choices.
