How to assimilate information accurately and why aortic dissections and PEs get missed- the pretest probability

Medical school and medical training teaches us that we do tests to confirm the presence or absence of disease. This is the wrong way to think about things. A better concept is to realise that we start with a certain pre-test probability of a disease, which is determined by the base rates of that disease in the population and the patient’s clinical history. Tests can only ever modify this pre-test probability into becoming more or less likely. At a certain point the disease may become so unlikely that testing for it causes more harm than good. This greater harm may come from radiation, reactions to things such as contrast dyes, harmful therapy that might be initiated as a result of a false positive result e.g. antibiotics for a blood culture result that is a contaminant, or simply the fact that time is wasted not pursuing the most likely diagnosis. Other times the disease remains so likely that you may have to pursue repeat testing (take for example the high false-negative rate of COVID swabs).

Consider this scenario. You are the on call house officer. You get paged to the ward to review a 35 year old patient who is having abdominal pain. He was admitted 6 hours ago with severe central chest pain that came on over a matter of seconds and lasted 2 hours. His troponins and ECG have been normal. He has now developed abdominal pain of the same severity and also reaching its peak over a matter of seconds. Concerned about the possibility of aortic dissection you look for mediastinal widening on the chest Xray, pulse defecits, or any neurological symptoms as you know these are the things to look for in a dissection. None of these things are present. Satisfied, you order further ECGs and troponins. The next day you find out he died overnight of an aortic dissection. The next day your consultant tells you “it just shows you how useless clinical exam findings are for aortic dissection- you can’t rely on them. Most dissections have a normal Xray!”

Is this correct? Are these clinical exam findings useless? Is the chest Xray normal in most dissections, as commonly quoted? Well, not quite. They are actually reasonably good tests, including the chest Xray (1,2). The problem is not taking into account the pre-test probability of an aortic dissection, which in this case is high based on the clinical history. Continue reading

The test is not the disease

When I was a trainee intern we had a patient on my general medical placement present with 2 days of right arm swelling and tenderness, with dilated superficial veins over her arm and upper chest. Her d-dimer was normal. She had an ultrasound of the upper limbs looking for a DVT. This was negative. With a negative d-dimer and USS we were all ready to discharge the patient home (who was otherwise well), however the consultant, a mentor of mine, insisted on a CT venogram. We all rolled our eyes. Eye rolling turned into eye widening as the scan showed the subclavian vein thrombosis we had all been missing. Continue reading

Avoiding harm: the early postoperative fever

One of the most important principles in medicine is to avoid doing harm to your patients. This is easier said than done because sometimes things that are iatrogenic are confused for natural evolutions of the disease process. That leads to writing the first post in a series entitled ‘avoiding harm’.

One good example is the early postoperative fever. I use this term to mean fever occurring in the initial 24-48 hour post operative period. As a house officer my friends and I spent many hours taking blood cultures on these patients, obtaining chest X-rays and chasing urine samples.

Eventually I realised what would happen is we would treat areas of atelectasis that were confused for pneumonia, colonised (but not infected) bladders and skin contaminants on blood culture. All of this represented unnecessary exposure to antibiotics, which have the potential to seriously harm patients.

Fortunately all of this can be avoided. There are 11 articles in the literature which I have taken the time to find for you. They uniformly tell us that if there is no sign of focal infection on clinical exam, the ‘septic screen’ can be safely forgone.

Blood cultures were the most useless. In four studies the pick up rate on blood cultures (excluding contaminant results) was zero. Two of these studies were designed specifically to look into the utility of blood cultures. All studies had large numbers of patients. One study found that of 38 blood cultures only 1 was positive and this was on post-operative day 16. Two further studies reported a 6-7% rate of positive blood cultures. The pick up rate of chest Xrays and urine cultures was about 10%.

Four studies reported that in those patients who were diagnosed with an infection, the majority of the time the source was identifiable based on physical exam and clinical picture, or that the clinical picture guided the need for further investigations.

Three studies actually attempted to quantify the cost. One calculated a cost of $8000 per change in clinical management, one worked out $2000 per infection diagnosed and one concluded that “rote” ordering of tests resulted in a total of $20000 (or $278 per patient) excess expenditure. All eleven studies concluded that “routine” ordering of investigations for early post-operative fever was unnecessary and costly.

  1. Sivakumar B, Vijaysegaran P, Ottley M, Crawford R, Coulter C. Blood cultures for evaluation of early postoperative fever after femoral neck fracture surgery.  J Orthop Surg (Hong Kong). 2012 Dec;20(3):336-40.
  2. Bindelglass DF, Pellegrino J. The role of blood cultures in the acute evaluation of postoperative fever in arthroplasty patients. J Arthroplasty. 2007 Aug;22(5):701-2.
  3. Lesperance R, Lehman R, Lesperance K, Cronk D, Martin M. Early postoperative fever and the “routine” fever work-up: results of a prospective study. J Surg Res. 2011 Nov;171(1):245-50. doi: 10.1016/j.jss.2010.03.009. Epub 2010 May 11.
  4. Fanning J, Neuhoff RA, Brewer JE, Castaneda T, Marcotte MP, Jacobson RL. Yield of postoperative fever evaluation. Prim Care Update Ob Gyns. 1998 Jul 1;5(4):146.
  5. Petretta R, McConkey M, Slobogean GP, Handel J, Broekhuyse HM. Incidence, risk factors, and diagnostic evaluation of postoperative fever in an orthopaedic trauma population. J Orthop Trauma. 2013 Oct;27(10):558-62.
  6. Ward DT, Hansen EN, Takemoto SK, Bozic KJ. Cost and effectiveness of postoperative fever diagnostic evaluation in total joint arthroplasty patients. J Arthroplasty. 2010 Sep;25(6 Suppl):43-8. doi: 10.1016/j.arth.2010.03.016. Epub 2010 May 10.
  7. de la Torre SH, Mandel L, Goff BA. Evaluation of postoperative fever: usefulness and cost-effectiveness of routine workup. Am J Obstet Gynecol. 2003 Jun;188(6):1642-7.
  8. Athanassious C, Samad A, Avery A, Cohen J, Chalnick D. Evaluation of fever in the immediate postoperative period in patients who underwent total joint arthroplasty. J Arthroplasty. 2011 Dec;26(8):1404-8. doi: 10.1016/j.arth.2011.02.019. Epub 2011 Apr 7
  9. Czaplicki AP, Borger JE, Politi JR, Chambers BT, Taylor BC. Evaluation of postoperative fever and leukocytosis in patients after total hip and knee arthroplasty. J Arthroplasty. 2011 Dec;26(8):1387-9. doi: 10.1016/j.arth.2010.12.024. Epub 2011 Feb 25.
  10. Verkkala K, Valtonen V, Järvinen A, Tolppanen EM. Fever, leucocytosis and C-reactive protein after open-heart surgery and their value in the diagnosis of postoperative infections. Thorac Cardiovasc Surg. 1987 Apr;35(2):78-82.
  11. Freischlag J, Busuttil RW. The value of postoperative fever evaluation. Surgery. 1983 Aug;94(2):358-63.

Why does hypoxia happen?

D0265FC6-2DF5-495D-86D1-D920BB9EF6C9
It is hard to avoid anything covid related in medicine at the moment. This pesky virus does raise an important point about pathophysiology however. There has been a collective surprise at the degree of hypoxia these patients can have despite a chest X-ray that doesn’t look that bad and at how comfortable these patients can appear despite their severe hypoxia.
Except there is nothing particularly new about this and it has long been a pitfall for the intern seeing hypoxic patients on the ward. Often one comes across a patient with post operative atelectasis (or viral pneumonia!) who has some opacities on chest X-ray (such as the one pictured above) but a degree of hypoxia seemingly unexplained by the chest X-ray. These patients are often sent for unnecessary CTPAs looking for PE.
Firstly, the sensitivity of chest X-ray for pneumonia and atelectasis is not as great as we think it is. Changes to the lung parenchyma are often more extensive than what is visualized on X-ray. Secondly the hypoxia is occurring on a microscopic level. Hypoxia occurs when there is mismatch between ventilation and perfusion i.e. blood flow to an alveolus is in excess of the lowered oxygen tension in that diseased alveolus. The larger volume of hypoxic blood that mixes with ‘good blood’, the worse it is. This is called shunting. Hypoxic pulmonary vasoconstriction is the reflex that protects against this mismatch but this reflex becomes less efficient with age, the presence of vasodilator drugs (pretty much antihypertensive or antianginal drug you care to mention), and a whole host of other physiological factors some of which may be specific to the disease itself.
Unwell septic patients with their high cardiac outputs will have a large volume of blood rushing through their pulmonary circulation which further decreases ventilation perfusion matching. The end result is that the degree of hypoxia is related much more to things occurring at microscopic level that we can’t see on an X-ray (on top of the fact that X-ray doesn’t tell us the true extent of parenchymal changes anyway). It is not uncommon to have post operative patients with a chest X-ray that looks like the one above but that are on 50-60% inspired oxygen.
These patients are tachypneic (because hypoxia contributes to the ventilatory drive) but they may not appear overly distressed because the work of breathing is more mediated by lung mechanics and the stiffness of the lung rather than the degree of hypoxia. If you’ve ever seen congenital cardiac babies with right to left cardiac shunts and resting saturations of 75% you’ll know what I mean. They can look blue but pretty happy.
So if you have a good reason for shunting (a high pre-test probability) such as being immediately postoperative or having a diagnosis of a viral pneumonia, that probability remains high despite what the chest X-ray might show you, and there is not necessarily a reason to go chasing a PE or invoking the presence of an unknown hemoglobinopathy (as many people are speculating with Covid). Of course, this doesn’t mean that looking for a PE is never indicated and clinical judgement in the individual situation is paramount. But you have to evaluate the patient in front of you, and not a computer screen, and if you understand the pathophysiology you’ll have a much better chance of doing this.
Till next time…

Tricky gut ischemia, the uselessness of lactate, and the importance of clinical suspicion

You are called to see Ms A, an 80 year old woman on the surgical ward, due to worsening abdominal pain and tachycardia. She was admitted 8 hours ago with the same abdominal pain and diarrhea and had a CT abdomen, which the radiology registrar has provisionally reported as showing non-specific pericolonic fat stranding. She has been treated as an infectious colitis. She has a history of ischemic heart disease, atrial fibrillation and claudication. Examination of her abdomen shows diffuse tenderness but no peritonism.

You call the surgical registrar to express your concern this lady might have ischemic gut. He informs you he is reassured by the CT findings, the lack of peritonism and the normal lactate, which you had decided to check because you have recently heard about the association between gut ischemia and elevated lactate. When you arrive at work the next morning you discover that overnight she had become septic, spiked her lactate to 8 and been taken for a laparotomy, where extensively necrotic bowel was found. She was palliated.

Ischemic gut is one of those diagnoses that is always tricky to make, as there is no lab test to help you and the examination findings can be non-specific, although “pain out of proportion to the exam” is what you might find in the textbooks.  Age confers an exponentially increasing risk, and past the age of 75 it becomes more likely than appendicitis or ruptured AAA (1). This is a fact which I certainly hadn’t appreciated and I suspect many people don’t, given the frequency with which we query the latter two on CT requests and the infrequency with which we query ischemic gut.

The first point to make abundantly clear is that peritonism is a late sign of extensive bowel necrosis so is not reassuring. The same applies to the finding of portal venous gas on an abdominal Xray, pictured below (2). The whole point is to diagnose the condition early enough that you can do something about it (either open or endovascular revacularisation). By the time these signs develop, the proverbial train has left the station.

portal venous gas

Continue reading

Bowel obstruction is a clinical diagnosis

You: “Hello Mr/Mrs. Surgical Reg, I’ve got a patient up here on the ward who’s started vomiting tonight and he hasn’t passed flatus for 24 hours and he’s got a tender tummy and I’m worried about a bowel obstruction”

Surgical Reg: “What does his abdo Xray show?”

You: “Its normal”

Surgical Reg: “Well why do you think he has a bowel obstruction then? The Xrays normal!”

Unfortunately when you create work for people, they can sometimes be less than helpful.

Now if you knew the sensitivity of an abdominal Xray, you could answer this question with ease.

This 2007 study (1) showed the sensitivity and specificity to both be about 82%. Second year radiology registrars had sensitivities as low as 59%, while senior radiologists reached up to 93%. Only 29 out of 90 patients had CT proven SBO. CT was the gold standard.

This study (2) used enteroclysis as the gold standard. This involves injecting a contrast material through a NJ tube and taking Xrays. This showed a sensitivity and specificity of 69% and 57%. If you are observant, you might have realised the implication of the numbers going down when a different gold standard is used- CT must not be so great either. Indeed, in this study the sensitivity and specificity of CT were only 64% and 79%.

Part of the explanation for these numbers is that Xray did better at identifying high grade obstructions (86% sensitivity) than low grade obstructions (56% sensitivity). The usefulness of the CT was in showing the cause of the obstruction rather than being far more sensitive.

In contrast (pun intended) this 1999 study (3) showed CT to have a sensitivity and specificity of 93% and 100%, while plain films were 77% sensitive and 50% specific. The gold standard was diagnosis from operative findings (25/30 patients) or by contrast study or clinical follow up, whatever that means, in the remainder.

Lastly this 1997 (4) review article is very informative and worth a read. It describes how the term ‘non-specific gas pattern’ means different things to different people! For example 65% of radiologists use the term to mean ‘probably normal’, 22% mean ‘can’t tell’ and the remainder mean ‘abnormal but not sure if ileus or mechanical obstruction’. What useful terminology!

They identify another analysis (page 1173, paragraph 2) where the sensitivity of plain films was only 66% when read by experienced radiologists; 21% of patients with ’normal’ findings had low grade obstruction.

So we can see that the accuracy of Xrays in the diagnosis of SBO varies a lot depending on the study, but to me seems generally underwhelming. At most the sensitivity in the hands of an experienced radiologist is 93% when compared to CT. It is important to remember however that the usefulness of the sensitivity/specificity depends on the pre-test probability- if your patient is vomiting and not farting and distended, the pre-test probability is high, therefore even with a high sensitivity there will be a large number of false negatives.

Clinical suspicion is important- radiography is a diagnostic aid, not the final arbiter. Always be guided by your clinical findings.

 

References:

  1. William M. Thompson, Ramsey K. Kilani, Benjamin B. Smith, John Thomas, Tracy A. Jaffe, David M. Delong and Erik K. Paulson. Accuracy of Abdominal Radiography in Acute Small-Bowel Obstruction: Does Reviewer Experience Matter? American Journal of Roentgenology. 2007;188: W233-W238. 10.2214/AJR.06.0817
  2. Reliability and role of plain film radiography and CT in the diagnosis of small-bowel obstruction. D D Maglinte, B L Reyes, B H Harmon, F M Kelvin, W W Turner, Jr, J E Hage, A C Ng, G T Chua and S N Gage. American Journal of Roentgenology. 1996;167: 1451-1455. 10.2214/ajr.167.6.8956576
  3. Comparative Evaluation of Plain Films, Ultrasound and CT in the Diagnosis of Intestinal Obstruction. Sudha Suri, S. Gupta, P. J. Sudhakar, N. K. Venkataramu, B. Sood & J. D. Wig. Acta Radiologica. Volume 40, 1999 – Issue 4. Pages 422-428
  4. The role of radiology in the diagnosis of small-bowel obstruction. D D Maglinte, E J Balthazar, F M Kelvin and A J Megibow. American Journal of Roentgenology. 1997;168: 1171-1180. 10.2214/ajr.168.5.9129407