Giving normal saline for SIADH

First question: will giving normal saline even help? Normal saline has an osmolality of 308 mEq (Na+Cl). In hypovolemic and some cases of euvolemic hyponatremia, it can help raise Na appropriately. If the urine osmolality is too high, giving normal saline will actually worsen hyponatremia because it will be like giving a more dilute fluid. 

Example: You have a man with suspected SIADH (he’s a smoker and has a new diagnosis of lung cancer). His Na is 120. His serum osm is 265 and his urine osm is 320. Your total free water loss is -37.5 mL, which is to say, if you give him 1 L NS, you will make zero difference. 

Let’s say the same man has a urine osm of 100. Your total free water loss is +2.08 L. If you give him 1 L NS, you will dramatically increase his Na. 

Let’s say the same man has a urine osm of 600. Now your total free water loss is -486 mL, which means if you give him 1 L NS, you will actually cause him to gain free water and worsen his Na. 

*It’s technically serum “osmolality” 
**D5 1/2NS osmolarity is often listed around 405, but it’s really less than that because dextrose moves more freely across cell membranes, and what I care about is the Na component

This chart is also very helpful for understanding how fluids are selected in common scenarios. 

Second question: How much normal saline should I give? The equation below is extrapolated from a calculation for how much hypertonic saline to give (see this post)

How much hypertonic saline should I give?

You have a patient with hyponatremia and you’ve determined that salt tabs and normal saline just won’t do, they need some 3% hypertonic saline. Great! You go to write the order…how many cc’s should they get? At what rate? Is there a better way of calculating it rather than just saying, give 50 cc x1 and hope for the best? 

As it turns out, giving 50 cc boluses and checking Na every 2 hours is an accepted form of empiric treatment. But if that makes you uneasy, there is a more mathematical solution. (Reference

You are solving for “predicted Na.” 513 refers to the amount of Na in 1 L 3% NS. Total body water is kg x0.5 for women, kg x0.6 for men

Example: a 60 kg woman with an Na of 108. Your goal is to increase her Na by 4 mmol over the next 4 hours. 

So far, your answer should be: the predicted change in Na will be 13.06 for every 1 L 3% NS. Therefore, to increase Na from 108 to 112 over four hours: 

=0.294 L or 294 mL. However, because you want to run this over 4 hours, 294/4=73.5 mL/hour. And that’s your hypertonic saline infusion rate!

Remember your Na correction goals:

  • 2-4 mEq in the first 2-4 hours (if you go over that, consider giving some free water back)
  • <8-10 mEq over 24 hours
  • You can stop being aggressive about correction around the time Na >125 mmol/L

What’s the deal with contrast-induced nephropathy?

How is contrast-induced nephropathy (CIN) diagnosed? 

Generally speaking, CIN is thought of as a more quickly reversible form of ATN: muddy brown casts and tubular epithelial cells can be seen in the urine. You should not see features of glomerulonephritis or AIN (RBC casts or WBC casts) or large urine output a la post-ATN diuresis. The surprise twist is that CIN is associated with a FeNa of <1%, which is more consistent with prerenal physiology.  Core IM podcast did a great episode on CIN that discusses that what we know about CIN is largely based on animal models, and so there is much about this condition that remains poorly understood. Importantly, they note that biopsy is not helpful because lesions in CIN are non-specific.

Should I give pre- and post-contrast hydration? 

Wyatt et al, in this commentary on the 2017 AMACING trial, a Dutch study looking at about 400 patients with CKD stage III getting contrast and risk of CIN, makes clear that the patient’s pretest probability for CIN, as well as the pretest probability for complications from fluid overload, matter a lot. Patients who have diabetes, hypertension, and obviously, CKD that borders on ESRD, are at higher risk. In addition, patients with ESRD who still make urine could be at risk, and giving contrast could worsen their renal function even more. Age (>60 years) may be associated but it’s not clear.if you’re working in an inpatient setting and have a patient with borderline renal function who is at risk of intra-op hypotension and low risk of flashing/pulmonary edema, it’s probably safer to give fluids. A 500 mL NS/LR bolus 30-60 minutes before/after is adequate.

On the other hand, there is very limited evidence to support yes fluids/no fluids. Previous studies that showed reduced risk of AKI with fluids might have shown this benefit because fluids prevented hypotension and ATN, rather than CIN itself. The PRESERVE trial, published in 2018, did not show differences in mortality, dialysis requirement, or persistent worsened kidney function, or CIN in patients receiving preventive IV sodium bicarbonate, oral acetylcysteine, normal saline, or placebo. But, there is always debate in the renal world…if you have a patient with CHF and an EF of 20%…giving fluids or sodium bicarbonate might not be a great idea.

Does the kind of contrast and procedure matter? 

Yes.Interventional studies, like coronary angiography, are higher-risk than diagnostic studies. In addition, according to the Core IM podcast, arterial contrast carries a higher risk of inducing CIN compared to venous contrast, although there are issues with differentiating CIN from atheroembolic showers, selection bias, and lack of control groups in these studies.



What are donor-specific antibodies and why do we test for them?

A quick refresher on HLA antibodies: Class I molecules are found on almost all nucleated cells. Class II molecules are found on antigen-presenting cells, B cells, and activated T cells. HLA matching is most important in bone marrow transplant and kidney transplant. In liver transplant, HLA matching may be beneficial but not so much that it is routinely performed; ditto for heart and lung transplant.

Screen Shot 2018-05-20 at 3.35.19 PM.png
Choo, 2007

Donor-specific antibodies (DSA) are HLA antibodies that are thought to contribute to rejection of the graft (allograft loss). They can be either class I or class II, although class II are more common.

DSA can be found pre-transplant (present in the recipient before transplantation occurs) or de novo (antibodies form in the recipient at any time after the transplant). Pre-transplant DSA are identified through panel reactive antibody (PRA). PRA is important because according to the OPTN, “Nearly a third of the OPTN renal transplant wait list is sensitized patients with a PRA of 10% or more. These candidates wait significantly longer than non-sensitized patients do and once transplanted suffer a greater risk of graft loss from rejection,” so these patients get additional “points” on the transplant list.

DSA can form at any time, although they are most commonly found <1 year after transplant. One study of kidney transplants found that after 4 years, 20% of patients with undetectable pre-transplant DSA will be found to have DSA, and that within 3 years, 25% of those patients will have graft failure. However, it must be noted that the mere presence of DSA does not equal rejection. Rejection is a diagnosis that must be made through pathology.

DSA in and of themselves do not need to be “treated,” although they are a risk factor for rejection. When and how to treat DSA is way above my paygrade, but suffice to say that if “DSA” are invoked, start thinking about the treatment of antibody-mediated rejection. There are a variety of strategies including ATG, upping doses of MMF or tacro, IVIG, and new monoclonal antibodies like bortezomib and eculizumab.



Can I give normal saline to a patient with hyponatremia who might have SIADH?

On the inpatient medical wards, not a month goes by without a case of a patient  with hyponatremia who gets 2 liters of normal saline, and when their sodium is rechecked, there’s no change or it’s even lower than it was before. Was giving normal saline the right thing to do?

If the patient has a good story for hypovolemia (has been taking twice the prescribed amount of diuretic, has not eaten in four days, etc.) then it makes sense that normal saline, with an osmolarity of 154 mEq/L, should help. It’s also a no-brainer that if a patient is grossly hypervolemic (has massive ascites, an acute heart failure exacerbation, etc.) then you should not give extra fluids, and diuresis will improve the sodium. The tricky one is “euvolemic hyponatremia,” and the most common cause of this is SIADH.

SIADH (syndrome of inappropriate ADH secretion) is a water problem. The body has too much water but continues to secrete ADH to retain free water (hence, inappropriate). One dictum you will hear is, “Don’t give normal saline to a patient with SIADH.” The argument is that in SIADH, salt handling is intact, so if you give a patient normal saline, all the sodium will be excreted but only half of the free water will be, which will make hyponatremia worse.

Is this true? Like so many things in medicine, it’s not an absolute rule. More important than whether someone has SIADH is how “saline responsive” they are likely to be. That is, will their sodium increase with normal saline or will it not?

How saline responsive might a patient be? The urine lytes are the key. 

  • urine osmolality: Uosm <100 is consistent with hypovolemia, whereas Uosm >200 is almost definitely not hypovolemia
  • urine Na: UNa <20 is considered consistent with hypovolemia (the body is trying to hold on on sodium to retain water) and >20 is considered fair game for a diagnosis of SIADH, although >50 seems to be the point of non-saline responsiveness. (paper here) UNa is unreliable in patients taking diuretics.
  • fractionated excretion of urea: FEurea >45% is considered to be non-saline responsive. (paper here)
UOsm <100, UNa<20
Malnutrition, beer potomania, surreptitious diuretic use, reset osmostat, hypovolemia already treated with normal saline
Uosm >300, UNa >40
SIADH, hypothyroidism, adrenal insufficiency

One review says that giving normal saline can unmask patients who are thought to be hypovolemic but actually have SIADH: “A rapid increase in FENa (>0.5% after 2 L of isotonic saline over 24 h), without correction of PNa, correlates with inappropriate ADH secretion.” This Hospitalist article recommends, “In salt depletion, plasma Na usually increases ≥5 mmol/L after 2 L saline infusion, which is not the case with SIADH.” So, a trial of normal saline can be diagnostic and therapeutic.  

The only patients you should NOT give normal saline to as a trial are those with an Na <120, as further driving down Na could put them at risk for seizures and coma. Patients with malnutrition or alcoholism and beer potomania are also at higher risk for overly rapid correction and osmotic demyelination, so use caution when repleting with fluids.

Want to calculate how much normal saline you should give someone? See this post. 

Can I give contrast to a patient with end stage renal disease or who’s on dialysis?

The short answer: almost certainly yes. And you don’t need to schedule urgent dialysis afterwards to remove the contrast.

Radiologists will tell you all the time that you can’t order a study with contrast because of someone’s creatinine, and so you may have to argue back, and advocate for your patient to get the study they need. I was inspired to write this post after our radiologist refused to perform a CTA on a patient on HD who turned out to have an actively bleeding 18-cm  hematoma. That’s a problem!

Think about it this way: a patient who is on dialysis does NOT have working kidneys, right? That’s why they’re on dialysis. Therefore, you do not have to protect their kidneys from contrast. The same is most likely true for patients on peritoneal dialysis, although I’m not sure if there are guidelines on this. There is an argument for preserving even residual renal function in ESRD patients, but…most of the time we are ordering CT scans for acute or urgent reasons, so you can do that risk-benefit assessment.

What you DO want to avoid is giving a patient on the cusp of needing dialysis, with CKD-IV or V, a huge contrast load. You don’t want to push them over the edge. Before giving them a CT torso with contrast, or sending them for cardiac catheterization, think about giving a little fluid beforehand and discussing with the renal experts. If it is a life/death situation, I’d probably lean towards giving contrast and hoping they don’t need dialysis permanently afterwards–but make sure there are no other alternatives and your consultants are on the same page as you.

What about MRI contrast? 

Most radiology departments have low-dose gadolinium protocols these days to avoid the feared nephrogenic systemic fibrosis (NSF). Anecdotally, nephrologists are more wary of approving gadolinium, and may be more likely to recommend dialysis after gadolinium load, but again, if there is a truly urgent need for an MRI study, the benefits likely outweigh the risks.

Do you need to schedule urgent dialysis after your patient gets contrast? 

Probably not, but the data is mixed. For example, this prospective study from NEJM suggests CVVH may help (in ICU patients), while other, small studies (here and very small study here) for HD. According to the American College of Radiology’s own 2016 guidelines:  “Unless an unusually large volume of contrast medium is administered or there is substantial underlying cardiac dysfunction, there is no need for urgent dialysis after intravascular iodinated contrast medium administration.”

Fun fact: a recent study from Hopkins in Annals of Emergency Medicine showed that contrast itself was not associated with higher incidence of AKI in patients getting CT scans, even in patients with CKD and on dialysis. (Of course, the doctors were more judicious with ordering CT scans in patients with CKD, and tended to give them fluids beforehand.)

How sensitive are muddy brown casts for ATN?

Screen Shot 2017-09-11 at 9.11.29 PM.png

Muddy brown casts are diagnostic for ATN, or acute tubular necrosis, which is an intrinsic form of AKI. But if you don’t see muddy brown casts, can ATN be ruled out? The literature on this is surprisingly light. There is one study of 267 patients with AKI who were classified as low or high pretest probability for ATN and had urine samples analyzed. The study concludes: “In patients with a low pretest probability of ATN (initial diagnosis of prerenal AKI), lack of casts or RTEC on urinary sediment examination had a sensitivity of 0.73 and specificity of 0.75 for a final diagnosis of prerenal AKI. The negative predictive value of lack of casts or RTEC in patients with low pretest probability of disease was 91%.” One important point to take away is that your pretest probability matters a lot for how confident you should be about making a diagnosis.

This review, by the same group, has a nice discussion section about making the diagnosis of ATN through a combination of methods, including microscopy, and use of FeNa and FeUrea.