Radiation Crunch Time in the ED by Stephen Colucciello, MD

Stephen Colucciello, MDThe Problem

Are you ordering too many CT scans in your ED?  Somebody is.  If it’s not you, it must be your partners who are the culprits.

I recently looked back at our ED CT numbers from 1999 and compared them to 2009. The numbers are astounding; a three fold increase in CT scans per year over the decade.  I can guarantee you that our volume did not increase by this amount.  I won’t even tell you the exact number of scans because my thyroid aches just to think about it.  From a national perspective, the total number of annual CT scans in the United States jumped from 3 million in 1980 to almost 70 million by 2007 (Arch Intern Med. 2009;169(22):2078-2086)

In one recent study, the authors estimate that approximately 29,000 future cancers may be caused by CT scans performed in the US in 2007 alone. (Arch Intern Med. 2009;169(22):2071-2077)  Others prognosticate that up to 2% of all cancers in Americans in the coming decades will be caused by CT scans.  (New England Journal of Medicine Volume 357:2277-2284  November 29, 2007  Number 22). 

Radiation risk from CT varies with age, sex, scan protocol and machine technology.  In one JAMA study, the authors examined the lifetime risk of cancer associated with radiation exposure from CT angiography (CTA) of the chest.  (JAMA 298(3):317, July 18, 2007).  They noted that the radiation exposure with triple-rule out CTA of the chest was up to 24 mSv; equal to approximately 250 chest x-rays.  They argue that the lifetime cancer risk for standard cardiac scans varies from 1 in 143 for a 20-year-old woman to 1 in 3261 for an 80-year-old man.  This begs the question, if the risk of cancer in a young woman is 1 in 140 for a single scan, do four scans over several years increase that risk to 1 out of 35?  Is the cumulative risk additive, geometric, or even worse?

As importantly, any CT we order on a patient today is unlikely to be their last.  In a study by Kline, 675 patients who underwent a CTA to R/O PE were followed for approximately 2000 days.  Seventy three percent had one or more subsequent CT scans of any body part and 5% had five or more repeated CT pulmonary angiograms.  (Ann Emerg Med, October 2008) 

Why the problem?

Not a huge mystery.  Just try to get a surgeon to see an abdominal pain patient without a CT (much less, take someone directly to the OR based on clinical exam).  Try calling a neurosurgeon without a CT.  Pan scans in trauma, CTs for broken heels, CT angios for chest pain, not to mention head CTs for children of nervous parents; including the kid who bonked heads with the family puppy.  The list goes on.

But beyond the value of the CT (which admittedly shows pathology pretty well), the demands of the consultants, patients, and parents, we selfishly use CT as the ultimate line of defense in the malpractice wars; “Hey, we got a CT scan, what more can we do?”  We know physicians will be held accountable for any missed pathology that could possibly be imagined on CT.  However, a plaintiff’s attorney is going to have a hard time proving that it was our particular scan that causes a thyroid cancer 25 years later (if we are even still around to sue).  Still, don’t we have some duty to protect patients against this malignant risk? 

Who should we worry about?

Children and young adults (especially young women) should be highest on our list; children because they are exquisitely sensitive to radiation, have years to develop radiation-induced cancers, and their future holds many emergency physicians eager for scans.  Young women are also at special risk, especially if the radiation is to the breast.  Obviously scanning during pregnancy (especially early pregnancy) presents a dual risk to mother and fetus.  On the other hand, octogenarians are unlikely to die of a radiation-induced malignancy that takes 20 years to develop.

What can we do?

Look before you scan; consider alternatives

Before ordering a non-emergent CT scan, quickly check your Radiology IS or electronic medical records for prior studies.  How many CT scans has the patient had in the past 2 years?  If the number is high (with multiple negative scans) consider if they really need another CT.  Would serial exams in the ED suffice?  If they do need some sort of imaging, are there alternatives such as US, MRI or plain films?  One straw poll of physicians suggested that one-third of all CT scans may either be unnecessary or could be replaced by non-ionizing imaging techniques, such as ultrasound or MRI.

Use Scoring Systems/Clinical Decision Rules

Patients with very low pretest probability for disease are unlikely to have a true positive test.  If you can identify the low-risk patient (less than 2% likelihood for serious disease), you can generally avoid performing a high-radiation imaging study.  In fact, low-risk patients generally need no additional testing.  The problem is getting, remembering, and using these clinical scores.  Carrying a clinical pocketbook in your lab coat, or using a PDA or smart phone loaded with the clinical decision rules simplifies the process.  (Maybe you will talk yourself into getting an iPhone after you read this blog!)  Read on for some useful scoring systems.

Work with Radiology Information Services to provide radiation alerts

It is relatively simple (as simple as things ever get with IS) to have the computer “count” the number of CT scans a patient has received in the last 3 to 5 years.  Whenever you or your secretary enters a CT order, an alert pops up on the order entry screen if the patient has had a pre-determined number of scans in your hospital system.  A more sophisticated solution (but still easy for the computer) is to account for the number of milliseverts for each type of scan (CTA of the chest yields more mSv than an ankle CT) and then incorporate a risk equation based on sex, age, and scan technique.  The hospital radiation safety officer can help with these calculations to come up with a alert.  The Emergency Physician may still elect to order the scan, but we might give pause if the patient has already had a million mSv earlier in the week……

Stand up to patients and their parents

It is always tough when the chief complaint is “I’m here for a CT scan of….” (The wrong response is; “would you like contrast and fries with your scan?”)  This demand usually relates to head scans for headache or minor trauma.  Instead of routinely “giving in”, make a big production with the neuro exam (the super-sized wobbly reflex hammer is sure to impress).  Say “hmmm” frequently after checking for Hoffmans’ sign and palmar graphesthesia (be sure to stroke your chin gravely).  After the exam, intone with relief, “Thank goodness you don’t need a CT scan.”  Female patients usually relent when warned about cancer risk from unnecessary scans; however this strategy will not work on men.  With insistent male patients, tell them the scan will shrink their testicles.

Strategies to reduce scans in particular circumstances

Pulmonary Embolism (PE) work up  

Scoring Systems and D-dimers

Use the PERC rule to determine who needs a D-dimer (http://www.mdcalc.com/perc-rule-for-pulmonary-embolism).  Low risk patients (low clinical risk by “gestalt” and negative PERC score) need no further work up.   A PERC negative patient is < 50 years old, has a HR < 100, O2 sat on room air > 94%, no prior history of DVT/PE, no recent trauma or surgery, no hemoptysis, no exogenous estrogen, and no clinical signs suggesting DVT.  The PERC rule identifies patients with such a low risk of PE (< 2%) that a D-dimer is not necessary; and thus eliminates the unnecessary CTA because of a false-positive D-dimer.

If the patient is not low-risk or is PERC positive, then order a D-Dimer (unless the patient is very high risk for PE; in which case go straight to imaging).  In all but high-risk patients (see Charlotte Rule http://www.medscape.com/viewarticle/575056_2), a negative d-dimer should end the work up.  As an added benefit, if aortic dissection is on the differential, a normal D-dimer will essentially rule out this diagnosis as well.

Bring back the VQ?

Consider getting VQ scan instead of CTA especially in young women suspected of PE but with a normal CXR. The radiation dose to the breast is about 0.28-0.9 mSv with V/Q scanning, but 20-60 mSv with 4-slice CT and 50-80 mSv with 64-slice CT.  An abnormal chest x-ray warrants CTA since VQ scans will usually be indeterminate in such cases.  This strategy can decrease the use of CTA for PE by 27% (Freeman, L.M., J Nucl Med 49(1):5, January 2008). 

Remember that patients with a low-probability VQ scan and high clinical suspicion, and all patients with an indeterminate or moderate probability scan will probably need additional imaging beyond the original VQ scan.  So in a few patients, this “VQ-first” strategy will increase their radiation exposure.

Suspected Appendicitis

“Ultrasound-First” policy (especially in children and women of childbearing age). 

If your surgeons insist upon an imaging study before evaluating patients with appendicitis, consider getting an ultrasound first (J Am Coll Surg 2009 Mar; 208:434).  If the US is positive for appendicitis call a surgeon, if the US is negative, re-evaluate the patient.  If you are still suspicious of appendicitis after the repeat exam, then get a CT. 

This strategy reduces ionizing radiation and relies on the fact that while US is very specific for appendicitis (> 95%) its sensitivity is less than that of CT (75% versus 95% for CT).  While “US-first” can be used in both children and adults, it is more widely studied in kids.

MANTRELS (Alvarado) Score

Start using a clinical scoring system to determine the need for imaging in adults, such as the MANTRELS Score.  Using this tool you can estimate the pretest probability of appendicitis and determine who can return to the ED for recheck, who should be consulted without imaging, and who should be scanned. 

Characteristic	Score
M = Migration of pain to the RLQ	1
A = Anorexia	1
N = Nausea and vomiting	1
T = Tenderness in RLQ	2
R = Rebound pain	1
E = Elevated temperature	1
L = Leukocytosis	2
S = Shift of WBC to the left	1
Total	10

A score below 5 is unusual with appendicitis, while a score of 7 or more strongly predicts of acute appendicitis (consider surgical consult without imaging).  In patients with an equivocal score of 5-6, CT scan (or a combination of US and CT) can reduce the rate of negative appendectomy.  "BestBets: The Alvarado Scoring System is an accurate diagnostic tool for appendicitis". http://www.bestbets.org/bets/bet.php?id=1671

Patients with scores less than 5 should be re-examined prior to discharge and if their score remains low they can return for a mandatory recheck in 8-12 hours (sooner if any worsening).

Recurrent renal colic

US +Plain film KUB Strategy

Some patients with simple renal colic receive a dozen or more CT scans of the urinary track per year.  Each time they present with flank pain, the emergency physician reflexively orders a CT without examining the medical record.  However, patients with classic recurrent renal colic may need no imaging at all, especially if previous imaging studies show the patient to have to have stones and no AAA.     If imaging is needed in a patient with multiple prior helical CT KUB studies, consider obtaining a plain film KUB in combination with a renal US.  This can often localize the stone and assess degree of obstruction (as well as rule out AAA).  (European Radiology Volume 14, Number 1 / January, 2004)

Adult Trauma

Pan Scans

Don’t instinctively order Pan Scans (CTs of head, c-spine, chest, abdomen, pelvis, and dorsal spine) on every trauma patient.  While these “stem to stern” scans can provide valuable information in the badly injured patient; not everyone needs a 40 mSv scan.  If the patient just needs a head and neck scan, perhaps a CXR, abdominal FAST exam, and serial clinical exams may be a sensible low-radiation alternative to pan scanning.  Consider performing serial FAST exams as an adjunct to the clinical exam in stable patients.

Clinical Decision Rules

Use clinical decision rules; Canadian C-spine rules (sorry Jerry) can decrease the number of cervical imaging compared to NEXUS rules (NEJM Volume 349:2510-2518  December 25, 2003).  There a variety of strategies to reduce the need for pelvis films, (J Am Coll Surg 2002 Feb; 194:121-5) and Chest CTs in trauma (J Trauma. 2007;62:631– 635.)

Pediatric Trauma

Pan Scans again

Stop the Pan Scans in kids.  After reviewing the number of Pan Scans ordered compared to number of unanticipated positive Pan Scan studies, we eliminated the “Pan Scan” as a single order set for children under 12.  If a poly-traumatized child clinically requires multiple CT scans, the physician must order each CT study separately (and not rely on a single “Pan Scan” order).  This resulted in a significant decrease in radiation exposure, with no known missed injuries to date.  

Use the PECARN Clinical Decision Rules. 

The PECARN, the Pediatric Emergency Care Applied Research Network has developed several valuable trauma decision rules that can decrease radiation exposure in children.

The PECARN head trauma rules (http://www.pecarn.org/publications/documents/Kuppermann_2009_The-Lancet.pdf) provide an evidence-base approach to neuro imaging in head-injured children.  The prediction rule for children younger than 2 years includes normal mental status, no scalp hematoma except frontal, no loss of consciousness or loss of consciousness for less than 5 seconds, non-severe injury mechanism, no palpable skull fracture, and acting normally according to the parents.  No child judged low-risk by these rules required neurosurgery.  PECARN also has similar rules for children older than 2 with head trauma.

PECARN abdominal trauma rules (J Pediatr 2009;154:912-7) is a 10 item system based on clinical examination, laboratory tests, and ultrasound to determine the need for abdominal CT scan and hospital admission.  The Blunt Abdominal Trauma Score (BATiC) includes: abnormal abdominal Doppler ultrasound (4 points), abdominal pain (2 points), peritoneal irritation (2 points), hemodynamic instability (2 points), aspartate aminotransferase >60 IU/L (2 points), alanine aminotransferase >25 IU/L (2 points), white blood cell count >9.5 g/L (1 point), LDH >330 IU/L (1 point), lipase >30 IU/L (1 point), and creatinine >50 g/L (1 point). A score of <7 had a NPV of 97% for intra-abdominal injury.  While this rule is a bit “clunky” to use in the trauma room, it’s a start.

Headache

Desperate for a CT

One article currently in press describes a woman with 58 head CT scans over several years (all negative).  She would go from ED to ED with classic symptoms of SAH or acute stroke; sometimes presenting twice in one week to different hospitals.  Her ability to mimic neurologic findings was uncanny (she was more convincing than most real stroke patients).  Her obsession with getting scanned amounts to a “Radio-Munchausen’s syndrome”.  However, at some point emergency physicians had to stop the madness and throw in an MRI or two.

Who Needs a CT for Headache?

Realistically, the answers to several simple questions can determine who needs an emergent CT scan for atraumatic headache.  Is the patient at risk for SAH (especially sudden-onset thunderclap headache that reached maximal onset within minutes to seconds)?  Do they have focal neurologic symptoms or findings?  Do they have any signs of depressed mentation or altered mental status?  Are they on Coumadin or otherwise coagulopathic?  Are they HIV positive or otherwise immunosuppressed?  Do they have a new type of headache after age 55?  Do they have a new type of headache associated with vomiting? (The Journal of Family Practice November 2005 Vol. 54, No. 11; AEM vol 4; Issue 7 Published Online: 29 Sep 2008

pg 654-661)

Future technology

In the near future, the CT scan may lose it preeminence due to advances in ultrasound and MRI.  3D ultrasound with contrast may someday replace CT scans for soft tissue imaging.  New high-speed MRI machines (a few minutes per scan) will probably become the standard for all neuro imaging ordered for headaches, TIAs, and stroke.  However, until that day, we need to restrain ourselves to some extent and use the CT in a more rational and radiation-sparing manner.

Table 1  From ARCH INTERN MED/VOL 169 (NO. 22), DEC 14/28, 2009

Summary

1. Don’t fry the young. 
2. Avoid unnecessary radiation to the female breast, especially under age 40
3. Use clinical scores.  Patients with very low pretest probability do not need high-energy imaging.  Use your smart phone to store your clinical decision rules and as importantly, use the rules in daily practice.
4. Check the computer or medical records.  See how many CT scans the patient has received in the past several years.  Talk to IS about computer alerts that pop up for patients with high radiation burdens.
5. Consider alternatives imaging (or no imaging at all).  Reasonable alternatives to CT include ultrasound, MRI, VQ scans, and plain films.  Getting one of these tests first may provide the answer you need.
6. Develop local consensus and clinical pathways; this way no one will be surprised with a VQ scan instead of a CTA on a young woman with R/O PE. 
7. Stay tuned for new technology; the 256-slice scan, multi-source scanners, radiation reduction protocols, and new computerized programs could rapidly the playing field.  Even better are the advances in non-ionizing radiation; 3 D ultrasound, US contrast protocols and high-speed MRI.

Comments

[Aaron Johnston]

I think complexities of the relationship between emergency room docs and other specialist groups is also part of the issue.
In Canada there is a definite radiology bias towards CT scan, particularly after-hours as the scans can be read remotely, while after hours ultrasound is frequently unavailable as many hospitals do not staff after hours US-techs (ie. the radiologist would need to attend in person). The pay scale for CT-scan reads also reflects a CT positive bias in terms of dollars per minute versus other modalities.
A different dynamic exists in the relationship between emerg docs and various speciality groups, and we now often 'need' a definitive diagnosis before asking specialist colleagues to evaluate the patient. We recognize that our colleagues are overworked, overburdened and we reflexively attempt to ameliorate this by only referring positive cases to them. In reality the patient with an atypical surgical abdomen may be better served by a physical exam by a surgeon than by a CT scan (or more commonly a second/third or fourth CT scan).
In my opinion the demanding patient scenario is a red herring, or at best a minor factor, in the entire argument. 'My child hit her head and needs a scan' can be interpreted as 'I am worried my child is badly hurt'. The tactic of addressing that worry rather than the CT scan defuses most of these situations. For the parents who remain worried, and who cannot be reliably coached to observe for red flags, a 24 hour observational admission is an alternative (though I have never had anyone take me up on that offer when the child is well and I have explained my concerns about future cancer).
My 2 cents..
Aaron

[Andrew Langsam]

Radiation from CTA's for PE in young women can be reduced by about 90% without degradation in the accuracy of the radiologists reading by applying a Bismuth Shield to the patients breast area. I routinely request a Bisthmuth Shielding for all of my female patients undergoing CTA.