In this Journal Jam we have Dr. Michelle Lin from Academic Life in EM interviewing two authors, Dr. Rebecca Smith‑Bindman, a radiologist, and Dr. Ralph Wang an EM physician both from USCF on their article “Ultrasonography versus Computed Tomography for suspected Nephrolithiasis” published in the New England Journal of Medicine in 2014. There is currently a wide practice variation in the imaging work-up of the patient who presents to the ED with a high suspicion for renal colic. On the one extreme, some EM physicians use CT to screen all patients who present with renal colic, while on the other extreme, other EM physicians do not use any imaging on any patient who has had previous imaging. The role of POCUS and radiology department ultrasound as an alternative to CT in the work up of renal colic has not been clearly defined in the ED setting. This study was a pragmatic multi-centre randomized control trial of patients in whom the primary diagnostic concern was renal colic, that tried to answer the question: is there a significant difference in the serious missed diagnosis rate, serious adverse events rate, pain, return visits, admissions to hospital, radiation dose and diagnostic accuracy if the EM provider chose POCUS, radiology department ultrasound or CT for their initial imaging modality of choice. This Journal Jam is peer reviewed by EMNerd’s Rory Spiegel.
Peer Review by Dr. Rory Spiegel (@EMNerd) May, 2015
Smith-Bindman et al have done a great job providing us with the largest prospective dataset of Emergency Department patients with suspected nephroliasis. Dr. Chan and Dr. Helman do a fantastic job deconstructing the trial and its intricacies, and Dr. Lin’s extra insight from the authors was especially unique and helpful.
The primary question posed is, “Can patients with suspected nephrolithiasis be managed with either point-of-care or radiology performed ultrasound rather than CT?” Though this is likely true, the Smith-Bindman et al trial is unable to sufficiently answer this question.
The trial’s pragmatic design led to a large portion of both the point-of-care ultrasound and radiology ultrasound groups undergoing CT scan during their initial Emergency Department presentation (40.7% and 27% respectively). In the authors’ interview with Dr. Lin they explain that this was likely due to the treating physician’s discomfort with ultrasound as an imaging modality. This may very well be true, but what we are left wondering is how many of these additional CT scans yielded further diagnostic information? Did the treating physicians order additional testing when the patient’s presentation offered a diagnostic dilemma? What pathology, if any, did this additional testing uncover? Did the Smith-Bindman et al trial find these 3 testing strategies equivalent because the additional CT scans performed in the ultrasound groups identified high-risk diagnoses, as defined by the authors’ primary endpoint, that otherwise would have been missed?
These questions speak to the primary measurement Smith-Bindman et al used to define effectiveness. The authors utilized the negative predictive values (NPV) of each respective diagnostic pathway to determine their efficacy in clinical practice. A low NPV can be accomplished in two fashions. First by using a highly sensitive diagnostic tool that identifies the majority of the patients with the disease in question. Alternatively if the prevalence of the disease is very low, then the NPV will appear low independent of the accuracy of the testing strategy. This distinction becomes important when the practicing clinician applies this diagnostic strategy to a population with a higher prevalence of disease than the cohort in which it was initially tested. In such cases the diagnostic sensitivity may be inadequate and the subsequent NPV will suffer. Unfortunately we are unable to determine the reason for the low NPV cited by Smith-Bindman et al. The authors do not provide us with the overall rate of high-risk diagnoses identified during the initial hospital visit. Without this important denominator we cannot assess if the low NPV is due to the sensitivity of ultrasound or the overall low prevalence of disease observed in their cohort. The point-of-care ultrasound and computer tomography groups missed 6 (0.7%) and 2 (0.2%) respectively. If these few misses accounted for the majority of high-risk diagnoses, it is likely this trial is underpowered to detect a difference between these diagnostic strategies. If applied to a cohort with a higher prevalence of disease, it is unclear how ultrasound would perform.
I think Dr. Helman & Dr. Chan summarized the take home points perfectly. The majority of patients presenting to the Emergency Department with signs & symptoms where the only diagnosis is nephrolithiasis will do well independent of what testing strategy they receive. In fact the majority likely require no imaging at all. The few that appear sick or present a diagnostic dilemma may still require a CT.
Smith-Bindman R. et al. Ultrasonography vs. CT for suspected nephrolithiasis. N Engl J Med. 2014;371:(26)2531. Full PDF
Yan JW, McLeod SL, Edmonds ML, Sedran RJ, Theakston KD. Normal renal sonogram identifies renal colic patients at low risk for urologic intervention: a prospective cohort study. CJEM. 2015;17:(1)38-45. Full PDF
Pernet J, Abergel S, Parra J, et al. Prevalence of alternative diagnoses in patients with suspected uncomplicated renal colic undergoing computed tomography: a prospective study. CJEM. 2015;17:(1)67-73. Full PDF
Other FOAMed Resources on Ultrasound vs CT for Renal Colic
For a Urology perspective that Dr. Lin refers to in the podcast, visit the Twitter Journal Club Urology JC – International Urology Journal Club #urojc @iurojc.
ER Cast podcast
St. Emylyn’s blog
The Skeptics Guide to EM podcast
Top 10 reasons NOT to order a CT scan for suspected renal colic on Academic Life in EM blog