Share on LinkedIn

CT_Scanner_01.jpg2d5efea2-a1b7-4c15-848e-4d6c5567eecfLargerExpertPeerReviewStamp2x200It appears that the excitement and utilization of computed tomography (CT) imaging in the emergency department (ED) has far outpaced our concern for the short- and long-term consequences of increased reliance on this technology. CT has greatly supplemented, or even replaced, our clinical decision making for many chief complaints.  Many articles document the dramatic increased CT use in contemporary practice, including a 330% increase in the rate of CT imaging from 1996 to 2007. The likelihood of a CT order being part of any ED encounter now approaches 15%, with no signs of decline (1). 

RISING TREND FOR CT IMAGING FOR RENAL COLIC

This trend especially holds true in the evaluation of flank pain (2). Kocher cited CT utilization rates for cases of suspected renal colic between the years 1996 to 2007 have increased from 4.0% to 42% (1).  This 10-fold increase in CT use occurred alongside a flat rate of urolithiasis diagnosis (18%) and hospital admissions for renal colic (11%) over the same period. During the same study period ultrasound (US) utilization was low to begin with, and only went down from 5% to 2.4% (3). Yet, the American College of Radiology appropriateness criteria rates ultrasound as equivalent to CT for the abdomen to evaluate recurrent stones. For acute renal colic in general they highlight an option to use US over CT unless patients fail to respond to conservative management or in those in whom surgery is anticipated. Why the disconnect between the evidence, guidelines, and our current practice patterns in imaging for renal colic? 
 

WHY CT IMAGING INDICATION CREEP?

 
Top10

TOP 10 LIST

Below are my top 10 reasons to think about when considering a CT in your next suspected renal colic patient:
  1. Cancer risk: An isolated CT scan is incredibly safe, but 1-3% of cancer is estimated to arise from medical imaging. A typical radiation dose for a CT of the abdomen and pelvis is in the range of threshold increased cancer risk among atomic bomb survivors; for a 40-year-old male the increased risk is on the order of 1 in 1000 (10).
  2. Patient expectations: Symptomatic kidney stones are common, and they usually recur.  Recurrence rates are estimated at 50% over a 10-year and 75% over 20-year period, with some people experiencing ten or more episodes over the course of a lifetime (11-12). If you set the expectation that all ED renal colic visits get a CT (analogous to all upper respiratory illnesses getting a prescription for antibiotics), it can be difficult to overcome this inertia on subsequent visits from a client satisfaction standpoint. 
  3. Imaging begets imaging: CT adds cost, time, and ionizing radiation exposure to the ED visit.  Malignant transformation risk from radiation is cumulative in patients who undergo multiple studies (13-14). 
  4. Generally good prognosis: Spontaneous stone passage rates are high, estimated at over 90%. One can further prognosticate based on size and location of stone, but this is only useful if having a complicated course of renal colic, defined by prolonged or uncontrolled symptoms with or without persistent hydronephrosis or signs and symptoms of an obstructing urinary tract infection (15). Clinically, complicated cases declare themselves.  In very high pre-test probability cases (good story and exam, positive hematuria) the CT is more likely to harm than help as incidental findings and cascade testing will outnumber serious alternative diagnoses (16).  
  5. Hydronephrosis is not an emergency: Hydronephrosis on acute presentation in the ED does not mandate an intervention. The evidence supports that hydronephrosis does not lead to irreversible loss of renal function if corrected within a few weeks (17). Ultrasound may not be as sensitive as CT for finding a patient’s kidney stone, but it is does very well at diagnosing hydronephrosis (18).
  6. Imaging begets urology referrals: Complications arise between 1-8% with urologic interventions (e.g. lithotripsy, ureteroscopy) and include sepsis, steinstrasse (stone street), stricture, ureteral injury, and UTI (15). Unnecessary CT ordering in uncomplicated cases might compel an early urology referral, which might lead to an early unnecessary intervention. The patient who would have likely done well with conservative management now gets a new trajectory, one that carries increased morbidity and cost.
  7. Wait and see: This is not just for acute otitis media (AOM) anymore. Patients are 3 times less likely to fill an antibiotic prescription when given a time- and action-specific contingency plan for when to fill a prescription given for AOM in the ED (19). Like AOM, renal colic is a largely self-limited clinical entity requiring supportive measures in the vast majority of cases.
Why not apply a similar approach to the CT order that many are doing now with antiobiotics for AOM?  If symptoms are controlled in the ED and the patient appears uncomplicated, they qualify for this algorithm. The timing of the “wait and see” decision might matter, although no researcher has examined this for renal colic. I have found that more patients will be accepting to defer a CT as their informed decision after their acute pain has been treated and resolved in the ED. There is evidence to show that patients with suspected renal colic who undergo delayed outpatient imaging have equivalent outcomes to those who undergo ED imaging at four-week follow-up (20). A systematic review of shared decision making in the ED shows it has been able to decrease resource utilization while increasing patient satisfaction, erasing the misperception that one always needs to order the maximum workup on a patient to ensure high patient satisfaction scores (21). 
  1. Reframing the question: Is finding the stone really the right clinical question to be asking? Instead, should we not be asking: what chance does my renal colic patient have of needing a urologic intervention, now or eventually? Because modern practice chases the former, we have largely left behind the ultrasound (US) in the diagnostic algorithm, since it often fails to visualize the culprit stone.  In a study of ED renal colic patients, 817 renal ultrasounds were examined for their predictive value of a urologic intervention. Of the 43% that resulted “normal” only 2 patients (0.6%) went on to a urologic intervention. That is, when the ultrasound is negative, you may still have a ureteral stone present somewhere, but it is highly likely to pass spontaneously. Of the 51% of patients with an US suggestive of or visualizing a stone, roughly 6.5% went on to a urologic procedure (22).  This further emphasizes the point that most renal colic cases are self-limited. It at least gives a framework to discuss options in real terms with the patient when deciding on an imaging strategy.
  2. UA + US = done: Emergency physicians have been shown to quickly learn how to accurately detect hydronephrosis at the bedside (18, 23) through an algorithmic approach. Hematuria on urinalysis (UA) is roughly 90% sensitive for a ureteral stone in acute flank pain. Ultrasonography (US) can increase the yield. In a study of 227, ED patients with acute flank pain, 122 were discharged without further investigation except urinalysis and bedside US. Ninety-nine had a urinary stone, and the others did not have a life threatening disorder (24). 
  3. Low dose CT: If you still want a CT, consider ordering it low dose. Niemann showed that radiation exposure can be reduced by roughly 10- to 20-fold without a significant compromise to sensitivity (97%) and specificity (95%) when compared to standard effect dose ranges (25). A variety of strategies and techniques exist to lower radiation exposure beyond simply not ordering the test. Protocols need to be developed between the ED and radiology departments for this to become a reality and not just a research novelty (26).
There will still be cases where the risk of benefit to harm is in favor of ordering a CT in suspected renal colic. Complicated renal colic occurs, but it usually declares itself clinically (uncontrollable pain and/or vomiting, signs and symptoms of an obstructing urinary tract infection). These cases will be the exception and not the rule. 
Some rare alternative diagnoses (e.g. malignancy, renal infarct) can present nearly identical to renal colic, but these are not especially time sensitive and will be discovered via a “wait and see” approach generally. 

CONCLUSION

To conclude, the vast majority of ED patients with acute renal colic do not need a CT. A urinalysis and bedside ultrasound will raise the pre-test probability adequately in most patients, if needed, and at least help screen for those who might go on to need an urologic intervention.  If a serious alternative diagnosis is really suspected or the patient fails to have their symptoms adequately controlled, a CT is reasonable. A low-dose CT is even more reasonable. Knowing they are receiving less radiation and possibly shortening their length of stay is likely to overcome any patient satisfaction lost by not ordering the CT.
Think before ordering that CT.
 

REFERENCES

 

Expert Peer Review

April 9, 2014
Before you read any further, be warned that I am of the opinion that we need to find alternatives to CT scan for the evaluation of patients presenting with suspected kidney stones and am in agreement with the top 10 reasons not to get a CT scan, and will not be discussing the contrarian view. Instead, I would like to make a few observations about how we arrived at this degree of CT utilization, as well as the direction we emergency physicians should take in order to safely decrease our use of CT going forward.
A close reading of the seminal articles by Smith that proved the superiority of CT scan over IVP reveals that CT scan was intended to be used in patients who would otherwise be referred for IVP. In a subsequent publication, Smith concluded that:
“Unenhanced CT is a valuable technique for examining patients with acute flank pain in whom a clinical diagnosis is uncertain” [1].
That seems to be a far cry from how CT scan has been used in the past decade, and the blog as well as many authors have described the phenomenon of indication creep. Aside from actual evidence based reasons for the increased use of CT scan, as well as patient expectations, other drivers of CT use likely include increased availability of CT scan, improved technology, economic incentives, and physician attitudes towards risk tolerance, uncertainty, and fear of malpractice.
Perhaps another reason for the rapid adoption of CT scan for the evaluation of patients with suspected kidney stone is explained by our metric for diagnostic tests – we commonly assign value to measures of sensitivity, specificity, predictive values, and likelihood ratios, rather than patient oriented clinical outcomes. As Daniel points out, there is little evidence to show improved patient outcomes for CT scan in patients with suspected kidney stone. In fact, our NHAMCS survey by Westphalen et al shows the opposite – despite a 10-fold increase in CT scan use between the years of 1996-2007, there was no change in the rate of kidney stone diagnosis, alternate diagnoses, or hospital admission. 
The evidence-based assessment of diagnostic tests should include patient oriented clinical outcomes. I am aware of potential solutions to the overuse of CT scan for suspected kidney stones on the horizon. One is the development of clinical prediction rules for kidney stone – Chris Moore recently reported the derivation and initial validation of a clinical prediction rule for the presence of kidney stone on CT scan [3]. While this rule will require external validation and impact analysis prior to broad adoption, it is a promising step in the development of a tool to help emergency physicians identify which patients require imaging and those who do not.
A second solution is ultrasound – both point of care ultrasound and radiology ultrasound. Stay tuned for the results of a multicenter randomized clinical trial comparing CT scan to ultrasound for patients with suspected kidney stone. It is possible that an algorithm combining a clinical rule and ultrasound could limit CT scan use to those who truly need it.”
References:
  1. Smith RC, Verga M, McCarthy S, Rosenfield AT. Diagnosis of acute flank pain: value of unenhanced helical CT. AJR Am J Roentgenol. 1996 Jan;166(1):97-101.
  2. Westphalen A, Hsia R, Maselli J, Wang R, Gonzalez R. Radiological imaging of patients with suspected urinary tract stones: national trends, diagnoses and predictors. Acad Emerg Med. 2011; 18: 700-707.
  3. Moore CL, Bomann S, Daniels B, Luty S, Molinaro A, Singh D, Gross CP. Derivation and validation of a clinical prediction rule for uncomplicated ureteral stone–the STONE score: retrospective and prospective observational cohort studies. BMJ. 2014 Mar 26;348:g2191. [free PDF of full article]
Ralph Wang, MD, UCSF Associate Professor of Emergency Medicine
Really great to see more conversation on this, and that EM is becoming actively involved in the discussion as well as the research. Read more about EM’s involvement in the Choosing Wisely campaign and development of our own lists for emergency medicine [JAMA Internal Medicine 2014 citation].