Welcome to this month’s edition of Staten Island Corner. The motivation for this comes from a common question that I am sure all of you have had at one point or another. The other day during a shift there was a debate about the accuracy of ruling in or ruling out DKA based on the shock panel alone. Everybody involved would make their point based on personal cases or the way the bicarbonate is measured, but when one of the attendings said, “What does the data show?”…..there was silence from everyone (including the attendings.) We all realized that nobody had any idea what the literature showed.
So here it goes…. From my search (for whatever that is worth) I found the grand total of ONE article that addresses this exact question. That’s right ONE. The authors of the one article also agree with me because they state in their introduction that they are not aware of any other studies addressing this question.
This prospective observational study was done in Los Angeles at USC. The goal of the study was to assess the accuracy of the VBG in diagnosing DKA and to look at the correlation between VBG and serum chemistry electrolytes in hyperglycemic patients. Patients over the age of 18 with a finger stick over 250 were eligible for the study. Exclusion criteria were critical illness, in police custody, unable to give informed consent, and VBG and serum chemistries collected greater than 30 min of each other. DKA was defined as glucose >= 250, anion gap > 10, bicarbonate < 18, pH <= 7.30, and presence of ketones on UA. 342 patients qualified for the study and 46 of them (13.5%) had DKA. The sensitivity and specificity of VBG for diagnosing DKA were 97.8% (95% CI 88.5% to 99.9%) and 100% (95% CI 98.8% to 100%). The PPV and NPV were 95% (95% CI 92.1% to 100%) and 99.7% (95% CI 98.1% to 100%). The positive and negative likelihood ratios were 582 (95% CI 36.5 to 9290.1) and 0.02 (95% CI 0.003 to 0.15). There was one case of DKA that was missed (bicarbonate of 19 on the VBG and 15 on the serum chemistry). The average difference between VBG and serum chemistry of the following were: Sodium by 2.0, Chloride by 4.0, bicarbonate by 1.6, and anion gap by 4.5.
There were two important limitations to the study that the authors acknowledged. The first being the small sample size and the second being the fact that other institutions may have other laboratory equipment.
The implications of this can potentially be huge in two important areas: total cost and ED throughput. If your only concern with a patient is to rule out DKA and you are able to confidently send one test (in addition to a UA for ketones) as opposed to multiple (we all know that if you are sending a chemistry panel, you will also send a CBC and probably PT/PTT) you will be able to save a significant amount of money on each patient. In addition to money saved, the turnaround time for a shock panel is significantly shorter than a chemistry panel. If we multiple this by the growing numbers of patients with uncontrolled diabetes in the US, we can potentially have a large impact.
I would like to hear what people think about this topic and how comfortable people are with just sending the shock panel.
References:
Menchine M, Probst MA, Agy C, Bach D, Arora S. Diagnostic accuracy of venous blood gas electrolytes for identifying diabetic ketoacidosis in the emergency department. Acad Emerg Med. 2011 Oct;18(10):1105-8. doi: 10.1111/j.1553-2712.2011.01158.x. Epub 2011 Sep 26.
basile
Latest posts by basile (see all)
- Staten Island Corner: VBG for DKA? - April 9, 2013
- Staten Island Corner: Coumadin Toxicity! - February 25, 2013
- Staten Island Corner: Video vs. Direct Laryngoscopy - January 9, 2013
4 Comments
andygrock · April 9, 2013 at 7:30 pm
I’ve had this same question/conversation after having THREE patients with mild or no acidosis on vbg with NO GAP end up having significant gaps on their cmp.
The literature shows you can trust the pH on the VBG.
I personally do not think a VBG gap is trustworthy.
The other question is does the gap matter.
ADA guidelines for DKA are gap, acidosis, low bicarb, ketones. If you have no acidosis on vbg is that enough for you?
Per this study, the average difference in bicarb is 1.6. Do we know the range of this though? Average alone isn’t a good statistical number for telling the accuracy between two values.
Do you folks trust bicarbs on VBG’s?
In the cases I had where the pt had mild to no acidosis and a gap, the attendings called dka and had me move pt to cct.
Ian deSouza · April 9, 2013 at 11:39 pm
NIce job, Basile.
I’ll raise your “only sending VBG+UA” and just go “only UA”. Remember most of the patients with asymptomatic hyperglycemia (i.e. sent from gyn, ophthamology et al. clinics to us) are just that – ASYMPTOMATIC. Therefore, one should have a very low suspicion (and along with that, the patient has a very low pre-test probability) of DKA. So, you so not need tests with a great sensitivity/NPV to achieve a good post-test probability. I believe in most cases a UA on its own will suffice, as it can be tested for ketonuria AND pyuria – with pyuria being a possible cause of the hyperglycemia which then will impact management.
Also, remember that a high anion gap in these patients can be due to other anions such as lactate (included in the VBG analysis) and unmeasured anions (as in acute renal insufficiency, which cannot); Both of these can confound a diagnosis of true DKA; .
nchristopher · April 15, 2013 at 8:01 pm
Thanks for the review Joe…
There’s a great reference card on this topic here:
http://academiclifeinem.blogspot.com/2013/01/pv-card-vbg-versus-abg.html
Personally I feel that making decisions off the pH on VBG is generally safe… You expect the venous pH to be slightly lower than arterial… so if a VBG is normal its reassuring. However, I have seen many cases where the gap is significantly different from VBG to CMP. The problem comes in when you are adding the error from each of the values that goes into calculating the gap (Na, Cl, HCO3). Based on the numbers Joe gave, at the extreme it could be a difference of almost 8!
The key thing to remember are there are a lot of causes of acidosis and anion gap… but you can only call it DKA in the presence of ketones! I agree that we shouldn’t even be sending bloods on many of these asymptomatic patients… but if you are sending them… send ketones as well! Also, if they have ketones and they are not hyperglycemic think about AKA and starvation ketosis.
TSmith · May 5, 2013 at 11:44 pm
Great discussion Joe! I agree that most asymptomatic patients only need UA (keeping in mind that not all ketones are picked up on UA); however, I do believe that’s all that is needed. Also, Joe keep in mind that shock panels cost more than a basic metabolic panel. I have heard somewhere between $150 – $250 for a shock panel. So you are not saving the patient. UA cost about $50.