CEimpact Podcast

Thiazide Diuretics and Hyponatremia

January 15, 2024
CEimpact Podcast
Thiazide Diuretics and Hyponatremia
Show Notes Transcript Chapter Markers

Hyponatremia is a side effect of most diuretics but the incidence from thiazides is unknown. Join host, Geoff Wall, as he evaluates a new study from Denmark evaluating hyponatremia and thiazides.
 
The GameChanger
Thiazide diuretics are considered first-line therapy but electrolyte side effects are underevaluated. A new study suggests hyponatremia is significantly more common than previously reported, especially in the elderly.
 
Host
Geoff Wall, PharmD, BCPS, FCCP, BCGP
Professor of Pharmacy Practice, Drake University
Internal Medicine/Critical Care, UnityPoint Health
 
Reference
Andersson NW, Wohlfahrt J, Feenstra B, et al. Cumulative Incidence of Thiazide-Induced Hyponatremia: A Population-Based Cohort Study. Ann Intern Med. 2023 Dec 19. doi: 10.7326/M23-1989. Epub ahead of print. PMID: 38109740.

https://www.acpjournals.org/doi/abs/10.7326/M23-1989?journalCode=aim

Pharmacist Members, REDEEM YOUR CPE HERE!
 
Not a member? Get a Pharmacist Membership & earn CE for GameChangers Podcast episodes! (30 mins/episode)


CPE Information
 
Learning Objectives
Upon successful completion of this knowledge-based activity, participants should be able to:
1. Discuss the use of thiazides in hypertension and their adverse effects
2. Assess the strengths and weaknesses of the Andersson et al study

0.05 CEU/0.5 Hr
UAN: 0107-0000-24-032-H01-P
Initial release date: 1/15/2024
Expiration date: 1/15/2025
Additional CPE details can be found here.

Follow CEimpact on Social Media:
LinkedIn
Instagram

Speaker 1:

Welcome to the Game Changers podcast, where we have clinical conversations that impact your pharmacy practice. Let's listen in as our team discusses this week's clinical practice game changer.

Speaker 2:

Hello and welcome to Game Changers clinical conversations. I'm your host, jeff Wall, professor of pharmacy practice at Drake University and internal medicine clinical pharmacist at Methodist Hospital here in Des Moines, and we'll just dive right into it. Today we're going to talk about a paper that was just published recently looking at the incidence of thiozide induced hyponatremia. Now, as we all know, thiozides you know, do you know? Because of their functionality they that's how they work is they work as a mild diuretic, and they work as a mild diuretic by increasing sodium excretion. Now they also have some effect on the brain and they have some effect on anti diuretic hormone release as well. So it isn't probably a surprise that that hyponatremia occurs in patients with thiozides. Certainly in the mid 2000s, when the all had study came out and for the more veteran listeners remembering that the all had study was at the time the largest study in hypertension. It was the one of the largest study whoever had hard outcomes. They just didn't look at the numbers of the blood pressure, but they also looked at rates of cardiovascular disease and kidney disease and eye problems and all that other stuff, and they found that that, all things being equal, thiozide diuretics were at least as good as ACE inhibitors and calcium channel blockers in preventing all these and actually some of them were better. So that caused a real shift toward using thiozides first line in in patients with uncontrolled high blood pressure. And that's fine.

Speaker 2:

I am certainly not not going to, you know, say that was a bad thing or anything like that. I mean I have certainly seen over the years is patients who come in who've recently started high to chlorothiazide for hypertension and they have had the bottom drop out of their serum sodium and they get real goofy. And this particularly happens in very elderly patients. And so the joke on my medicine services you know starting, you know, high to chlorothiazide on an 85 year old little lady who already has a relatively low sodium to begin with. Because, remember, as we all age our osmostat changes and we are certain sodium just start to march down and again, you know, if all else kind of stays the same, that's not that big of a deal.

Speaker 2:

Your body gets used to a sodium of you know 128, 129, 130. But then we start the patient on a thiozide diuretic and the bottom just drops out of their, their sodium because they can't afford to lose, you know, three or four points of sodium and then they start getting into the danger zone of, you know, confusion and dizziness and you know, even as bad as seizures and things like that. So you know certainly my clinical gestalt over the years has been, you know I can count on my fingers toes multiple times when we've had this exact case where you know a very elderly patient started on on a thiozide diuretic for high blood pressure control, ends up in trouble because of low sodium. Now a completely different argument is should you be starting brand new anti-impertensives on an 85 year old anyway? My personal opinions probably know. But again, that's another podcast for another time.

Speaker 2:

So we're really going to take a look at the overall incidence of hypotremia that was examined in a large population core study from the Netherlands. Of course, as we always say on this program, one of the advantages of having a centralized or socialized medicine program is you always have really, really good pharmacovigilance programs because you have basically one insurer, you have basically one pharmacy group that dispenses everything and you can get all that information from those. So that that's why a lot of these studies tend to be done either in Australia or in Western Europe. Again, the study done in the Netherlands. The reason they wanted to do it is, again, you know, in the study very clearly state that they're not trying to say that the eyes I diuretic should be used or that they don't have good outcomes in patients with with high blood pressure. But like, I think, a lot of people, including myself, they noted that that when you take a look at the package insert of diuretics, they call hypotremia rare and the rare definition is anywhere from less than 100 to less than one in 10,000, which is a pretty widespread in my opinion. But I mean, that's usually what's considered rare in package inserts and that just didn't really jive with, I think, what what a lot of these practitioners were seeing.

Speaker 2:

There isn't a whole lot of other data on that. I mean, again, with drugs as old as this is, you'd think there'd be several studies that took a look at at hypotremia in these patients and there really is. A couple of small studies have found cumulative incidences over time, usually in the kind of the three to five year range. So patients who had been on a thighs sites direct for three to five years would find an overall incidence of anywhere from three to 5%, so again, far higher than the less than a one in 100, right. And recently a retrospective court study found it was even more common than that, but it was based on a relatively small number of patients. So in the bottom line is that we had a lot of single arm retrospective studies or poorly done small court studies that had tons of confounders and things like that.

Speaker 2:

That really didn't answer the question really, what is the incidence of hypotremia? And the other big problem of these studies is they didn't really talk about severe hypotremia. Again, you know, someone can have a sodium of 130 and that might be considered hypotremia in the lab of your hospital, but most patients are going to have serious symptoms from that, whereas someone who has a sodium 120, a lot of patients are going to have symptoms of this. So that's kind of the rationale for wanting to do this study. As I said, it was a Netherland study where they took a look at the dangerous registers and again, like all of these Western European studies, they have, you know, just this unbelievable database of all this information on patients and it includes prescription, drug use, laboratory results. You know the individual electronic patient records and nodes, etc, etc, etc, all the diagnostic codes all basically in one large group of databases so they don't have to go or look into different databases.

Speaker 2:

This was a pragmatic target study which I thought was pretty good, and they wanted to use observational analysis to emulate pragmatic target trials, eliminating the increase of or estimating the increase of the cumulative instruments of hyponetremia, and I think that's another key piece of this is they wanted to all about just look in the first six months, but they wanted to look downstream, several years downstream and see if the incidence of hyponetremia increased. And the other nice thing about this study is they actually had controls. Now, again, this is an Arama's control trial, so keep that in mind. But they actually had kind of a case control setup because they compared patients in this database who received thiazide diuretics or non thiazide diuretics and they actually had kind of a two by two design where the first group they looked at patients who got a single thiazide diuretic and in this case the thiazide diuretic is one we don't use in the United States Bendro fluoro methozide never seen that used in the US, but it is a thiazide diuretic compared to calcium channel blockers. So this first arm was either you were on the thiazide or on a calcium channel blocker. And then the second part of the two by two group is they looked in patients who received a combo of hydrochlorothiazide and either an ace or an arb, versus the ace and arb by itself.

Speaker 2:

Apparently, in the Netherlands hydrochlorothiazide is almost never prescribed by itself. It's usually prescribed as a combination, as we often do it here in the United States, and other thiazide, such as this Bendro flu methozide, are more likely to be prescribed separately, just for high blood pressure. So I love that drug, the Bendro flu methozide. It has a potency I looked this up fairly similar to hydrochlorothiazide, and the doses are really similar and stuff like that. So I think I think it's reasonable to extend what we would know about that drug compared to hydrochlorothiazide or chlorothaladone here in the United States.

Speaker 2:

Basically, so they know that in Denmark all of these drugs are considered first line choices in their national guidelines for uncontrolled hypertension. So you know they expected to see quite a few people on this, these various and sundry complications. One of the things they spent a lot of time talking about is, though, calcium channel blockers have not really been reported to cause hyponatremia. Aces and arbs have, though again, it's a very rare incidence, at least according to the package insert, and so they did try to account for that in their study design.

Speaker 2:

And I think it's one of the reasons why they had these kind of two by two factorial things to kind of look at at combination therapy versus single therapy as well. So I think you know this two by two design actually was pretty clever because it allowed them to kind of deal with a lot of confounders that you'd have a difficult time dealing with. Basically, so, eligibility criteria just had to be over age 40. You had to receive a anti-advertensive between 2014 and 2018, with no use of anti-advertensive drugs within the past year, and you had to be able to be found and lived in Denmark and have all your data there, etc. Etc. Etc. They did not actually interesting require that a diagnosis of hypertension be in the patient's electronic medical record. Kind of interesting, so, and they did this for a couple reasons. One, they actually had access to only the diagnosis and the secondary health care you know part, so they thought they may be missing people. And they also pointed out they small, but not inc. Considerable number of patients take thiozide diuretics for other reasons. Some people still even though I never recommend this take thiozides for patients who have pedal edema as they get older. I wholeheartedly go against that, but it is not uncommonly done. Thiozides are also occasionally used for, you know, kidney stones and things along those lines. So you know, I could see why there's a number of reasons why they did not want to absolutely have a diagnosis for an inclusion criteria.

Speaker 2:

However, they did look at patients who had hypertension in their list. They just knew that it may not be complete. They looked at patients who were hospitalized. They looked at patients who had various sensory other disease states like heart failure or hepatic impairment or kidney impairment, which all, of course, in and of themselves may lead to hypoenetremia. So they tried to do that. They looked at patients who had a diagnosis of diabetes and semitis. They looked at patients who had a diagnosis of SIDH again, all things that might lead to this and excluded most of those patients. So basically tried to get a pragmatic group of patients and I think I would see patients who have multiple disease states, most of whom are on thiozides for hypertension but may be on thiozides for other reasons, and that included patients with diabetes and some degree of renal insufficiency and things along those lines. So when that was all done, they obtained information about field prescriptions from the Register of Medical Product Statistics, which is again this big pharmacovigilance thing that every time you get a prescription filled at a Danish pharmacy, it includes just about every piece of information about that prescription date, strength, et cetera, et cetera. So again, they were able to take a look at pharmacy-based level data pretty easily.

Speaker 2:

The primary outcome of interest was hypoenetremia. Big surprise there. They defined that as a laboratory test level of less than 130 millimoles per liter. That was a primary outcome. However, they also looked at severe hypoenetremia, which they counted as a sony muslin 125. And that's certainly where I would start to think you're going to start getting into trouble Hospitalization with hypoenetremia and any hypoenetremia less than 135. And they also looked at hypernatremia as a negative control outcome. They did want to look at all-cause mortalities as a secondary outcome, but the numbers were pretty low and I think, even in their own analysis, that they probably did not have enough patients to really show that.

Speaker 2:

So they then did estimates, basically estimated groups instead of just estimated numbers of interest, and they did do an intention to treat analysis. All these incidences were using the Kaplan-Meier estimator and they had a very complex system where they tried to account for a ton of covariance, which again is kind of all the things you would think of age and gender and other medications they were on and how long they had hypertension. You could go on and on. As a model that was pretty complex and they used Cox regression for that. So, as far as the group themselves, the source cohorts of the gigantic database they were looking at, they had a cohort of 2,300,000 patients age over 40 who were initially naive to anti-advertisement drugs, starting in 2014. Of that, 2.3,297,000. Then between 2014 and 2018, we're started on an hyper and hyper-intensive drug either the thiazide, or calcium channel blocker, or the combo thiazide, ace-nibber, arb, or just the ARB and ACE-nibber by itself. So about 300,000 there.

Speaker 2:

In the first cohort they excluded about 42,000 patients for a variety of reasons, and then that was the health and channel versus the thioside arm, and then in the other arm they excluded about 73,000 patients. The primary one for that was that a lot of them actually had a higher incidence of hypernetic tremia than they thought and that they had a difficult time because of switching. Ok, where you're on a thioside, then you're on a thioside, an ACE, then you're on a combo. They had a hard time switching. They're teasing that data out about OK, what were you on and how many of these patients switched back and forth, so they ended up having to exclude a lot of those patients.

Speaker 2:

So, when it was all said and done, they had about 83,000 patients in the thioside versus calcium channel arm and about 97,000 patients in the combo ACE or ARB versus just ARB group themselves. They censored a number of those patients. When it was all said and done, a lot of them died or emigrated or had some other reason for having their data censored. And so, unfortunately, when it was all said and done, even though we started with about 2.4 million patients, we ended up with, I mean again, big numbers. I mean there's no doubt about that, but the numbers were smaller than 2.3 million and you kind of end up with the number of events being relatively small across the board. So what did they find then when they looked at all these patients and what were some of the backgroundies of those patients? So we'll answer that question right after we hear from our sponsor, ce Impact.

Speaker 1:

Are you a pharmacist by design? Since we hold a vital position on the health care team, it is our responsibility to advance our knowledge and skills so we can provide the best possible care to our patients. Being a pharmacist by design means striving to be the best version of ourselves, not just as professionals, but as individuals dedicated to improving patient outcomes. Learn more about pharmacists by design at ceimpactcom. Join us and begin your journey to being the best version of your pharmacist self.

Speaker 2:

So we are back talking about thiazide diuretics and the risk of hypoenetremia in a large cohort study that was done in the Netherlands, taking a look at these patients. Again we take a look at the total numbers of patients. They were much smaller than the 2 million or even the 400,000. But again they had some numbers associated with it. So in the thiazide versus calcium channel blocker use there was about 38,000 patients in the thiazide versus about 45,000 in the calcium channel blocker group. And then in the other part of the two by two analysis where they had hydrochlorothiazide combined with either ACEs or ARBs, that was about 11,000 patients where ACEs and ARBs by themselves was 85,000 patients.

Speaker 2:

So kind of interesting, female gender was relatively similar, except interestingly and again I think this may be one of the reasons that I've seen this a lot they had that 74% of patients who received the thiazide diuretic by themselves were female. So that number was far higher, especially for a 50-50 split in every other arm. So the calcium channel blocker arm, the ACE inhibitor ARB arm and the ACE-ARB thiazide combo arm were all more around about 50% male, female. The mean age was about 63 across the board.

Speaker 2:

So again, I'd say not super duper elderly but I would say kind of middle-aged, and when you take a look at the curves of the number of patients, it really had only about seven to 10 patient percent of patients across the board who were over age 80. So kind of keep that in mind as we're talking here. When they were looking at their baseline laboratory tests, of that very first test they got in that 2014 to 2018 range, they about only about 60% of patients even got a baseline serum sodium measurement, which I guess is probably not that unusual. And then they found that the means in the ones where they got that means sodium level was smack dab or it needed to be in as 140. Mean potassium was smack dab or needed to be at 4.0. And a very, very small percentage of patients in fact less than 5% in any group had a EGFR of less than 60 mils of minutes. So these, all these patients had had really good renal function. The other interesting piece is that every small percentage of these patients had hypertension. Again, less than 2% of patients had hypertension. So much more surprising and, again, probably not very applicable to the patients that I see here here in the US. I mean taking a picture of all that, these, you know, on average. You know again, this is on average. These were pretty well patients. These were patients who are kind of middle aged, doesn't seem to have a lot of other issues or difficulties. Small percentage of cardiovascular disease, you know again. Small percentage of patients with diabetes or cancer, apparently no significant renal dysfunction. So for all those reasons I mean again, the morbidity cohort of their level with these patients was, I think, a little lower than I would tend to see.

Speaker 2:

They did take a look at individual medications that may have a role in causing hyponatremia, particular nonsteroidals and, as you might imagine, since they're over the counter, about one quarter of patients reported using even occasional nonsteroidals and the end-to-end lead type on atremia as well. Opioids interestingly about 12% of patients, 15% of patients reported using them. Antipsychotics was about 4% and agents for COPD was about 10%. So kind of interesting there. So then we get to the meat of the results. Among the individual thiazide initiators, about 70% of patients received a first prescription for a dose of 2.5 milligrams of the Ben-Rosthura methozide and 2.5 is equivalent to 25 milligrams of hydrochlorothiazide and about 30% of patients received 1.25 milligrams, which is it converts to 12.5 of hydrochlorothiazide.

Speaker 1:

So again.

Speaker 2:

I think that's about where you'd see most patients in the United States started on thiazides, either 12.5 or 25. Remember that there's really almost never a reason to go above that for really any indication that certainly hypertension, you just get more side effects. This did not differ materially across subgroups. Then when they looked at different ages and genders and other things, there really wasn't a big difference there. They also found that adherence, at least as far as pharmacy records, was about 80%, which is, I think, pretty good, and it was equal between the thiazide versus calcium channel blocker use. Interestingly, slightly better for the hydrochlorothiazide combo with the ACE or R compared to the ACE by itself. Again, probably all that surprising, because you're only taking one pill a day versus two pills a day, as you might think.

Speaker 2:

There, baseline levels, as we noted, were of sodium. Sodium levels were very similar between the use and their follow-up and all that seemed to be pretty similar between the use as well. So what did they find? Well, in the first group of the thiazide versus calcium channel blocker, the two-year cumulative incidence of hyponatremia for a serum sodium but less than 130, was 4% compared to 1.35%. So, and that was statistically significant, and so basically a doubling of the incidence of hyponatremia in a two years following the initiation of the drug. So again, that's far, far higher than the less than one in 100. Right, so I mean?

Speaker 2:

now you're talking three in 100, almost four in 100 of patients who might develop hyponatremia, you know, if they're on a thiazide versus calcium channel blocker. Interestingly, the numbers were very much the same for the combo of the hydrochlorothiazide R ACE or just the R ACE by themselves, with again 3.5% two-year incidence of hyponatremia versus 1.35%, which again was statistically significant. So I think there was some thought that that number would be higher, be closer to each other, because again, aces and R's have been reported to cause hyponatremia. But that really wasn't the case. When they looked at the Kaplan-Meier plot, as you might expect, that the hazard ratio for hyponatremia was highest in the first months makes sense, but it did. Once it reached a threshold, low threshold it did not decrease any further. So it was absolutely possible for people to develop hyponatremia as the two-year follow-up period went on and again we don't know why that is maybe somebody became dehydrated or something along those lines.

Speaker 2:

But the bottom line was that even though the risk was higher and the hazard did decrease, it did not go back to the baseline level that the comparator arms were. So I think that's kind of worth noting. When they took a look at severe hyponatremia, the incidences were lower, but the pattern was pretty much the same. It was about one and a half percent of patients in the thighs I had over two years who developed severe hyponatremia, so in less than 125 compared to, again, that 3% of patients. So again, about 1.5% of patients might develop severe hyponatremia compared to the compared drugs. Basically, the other thing they did notice when they did their analyses was and again this kind of always nice when data confirms your own biases they found that the overall risk of severe hyponatremia increased dramatically after age eight, so patients were over age 80, that incidence went much higher and was actually more closer to 6% in patients who had hyponatremia and about 3% of patients who had severe hyponatremia. So basically it almost doubled the incidence of the risk in patients over age 80. Hospitalization for hyponatremia was also statistically higher as well, again, especially in patients over age 80. Those numbers weren't as high in patients under age 80 and it did not reach statistical significance.

Speaker 2:

So they in their discussion they basically say that this population-based cohort study that used observational data, they found a substantial risk access for hyponatremia with initiation of treatment of thiozidiuretics compared to other first-line anti-abertensive drugs and they basically note that it was much higher than the less than 100, the less than one in 10,000 rates that are commonly quoted in the package insert. They note that again, the risk of severe hyponatremia was higher in older patients and in patients who are on multiple medications. Again, that kind of stands to reason and this follow-up, even though it dropped over the first few months, stayed consistent and there was a statistically significant difference between that and the controlled anti-abertensive throughout the two years of treatment. So they note that this is much higher than had been thought of previously and again, certainly it's not in line with the package insert.

Speaker 2:

Reading of that and while on a par with some of the other retrospective studies have done Ashley, was a little bit higher than I think some of the other studies have found as well, because a lot of these they were pushing more like 4% to 5%, especially in patients who were over age 80. And then all these studies do a large section on some of the limitations of the study. Again, it is a retrospective course study. We're not going to be able to do ever a randomized control trial that just looks at side effects. So this is probably the best data that we're probably ever going to get.

Speaker 2:

They note that they only looked at these two thiosites. It's entirely possible that other thiosites and they actually called out a chlorthalodone isn't really used in Denmark, so the generalizability to those drugs is unclear. Knowing the pharmacology's medications, I have no reason to believe there would be a whole lot of difference between them, but that is something that they note. They note that this was their only outcome looking at hypotension. They did not look at blood pressure control. They didn't look at hard outcomes associated with hypertension. This was designed purely to take a look at the safety of the drugs, so it really can't make any generalizations based on that. They note that the overwhelming number of patients who were being seen were by general practitioners, and so is it possible that if somebody's hypertension is being followed by another health care professional, the numbers might be different.

Speaker 2:

Again, my personal opinion, at least here in the US, is I'm not sure it would be any different, depending on anybody who prescribes them, but that is worth noting. They note that they weren't able to capture everybody using over-the-counter medications such as nonscroydals. That may have played a role, and of course they tried to take care of all the confounders, but you're never able to do that in the study. They do note that ascertainment bias was probably their biggest concern and that to mitigate that they wanted to make sure that this wasn't hyponatremia with a sodium of 132 or 133. That I think they said I think reasonable cutoff, I swear they would consider hypernatrinolusal 130 and severe hypernatrinolusal 125. And again, I think those are clinically relevant outcomes, even though you know I don't see a lot of people seizing with serum sodium 130. Yeah, I'd say once you drop below 130, it definitely causes most clinicians to kind of be a little bit concerned about what's going on. They of course weren't able to measure, you know exact measure, adherence or anything along those lines.

Speaker 2:

So again, those are all things that happened with almost all of these kinds of studies. But what I kind of took away from the paper and again, you know, yeah, it's always nice when a study confirms your bias is that you know. Like you know, I'm not trying to say that and I don't think these guys are trying to say that the other side should not be used and shouldn't even be used. It should not be used in the elderly. I think that hypernatrimia, especially in patients over age 80, I think you really need to take a very, very hard look at risk versus benefit. And you know, I would argue, in a nonambulatory patient or a patient who is not living independently, you may not need to control their blood pressure all that well anyway, but if you are going to pick one, I think that in many of those cases the harm associated with putting them on a thiozide diuretic is going to outweigh their benefit, because you know it's going to take years for those drugs to show difference in cardiovascular or renal outcomes and it only takes a couple of months for their sodium to blot them out and they get goofy and fall and hit their head. And so you know, I would argue that if you plan on treating somebody with high blood pressure, especially in the very elderly, that thiozides may not be the first drug you reach for.

Speaker 2:

I think that we have a wide variety of other drugs, particularly in calcium channel blockers, aces, arbs. They don't seem to cause this problem and I think in appropriate patients, you know not saying that thiozides should never be used in those patients, but I think that this study, which is something I've been telling my internist for years now, you know, I think adds more meat to the recommendation that thiozides really should probably not be first line in the very elderly, especially if they're not completely independent and not having any other, you know, issues out of dementia and along those lines, and that if you do start thiozides on these patients, that at least in that first two months grabbing a serum sodium is probably a good idea. So that's it for this week's edition of Game Changers. Hope you like what you heard.

Speaker 1:

We will see you next week Until then remember time flies.

Speaker 2:

I don't know where it's going, but the most important day is today. We'll see you then. We'll see you then.

Speaker 1:

Jen here. Be sure to check out our education at cempackcom. You'll find it to be your one stop shop for all the CE resources you need. Become a pharmacist by design member today to access it all for free, including CE for this podcast. Thanks for listening. We'll talk to you next week on Game Changers Clinical Conversations.

Thiazide-Induced Hyponatremia Incidence Study
Thiazide Diuretics and Risk of Hyponatremia