My understanding is that shortly before my twin and I were born, a nurse used a pencil test to predict our genders. Though my mother is quite smart, she believed the nurse knew what she was doing. However, shortly thereafter, it turned out that I had a twin brother not a twin sister.
ER doctors often perform a similar service to the pencil test when they use an abdominal radiograph (AXR) to determine if their patients have constipation. A new pediatric study from Toronto highlights this phenomenon and current recommendations (J Pediatr 2014; 164: 83-8).
Background: this retrospective cohort study of children <18 years took place between 2008-2010. As part of the study, a single pediatric radiologist (blinded to participant classification, assigned Leech scores to all misdiagnosis AXRs along with 20% of the remaining AXRs. From a total of 112,381 ER visits, the review identified 3987 where constipation was the discharge diagnosis (3.5% of all visits). In the cohort diagnosed with constipation, the mean age was 6.6 years.
- Only 9% of children returned within 7 days. 20 of these (0.5%) had a significant misdiagnosis based on the authors definition, including 7 with perforated appendix, 2 with intussception, and 2 with bowel obstruction. Other misdiagnosis included ovarian torsion, thalamic brain tumor, acute lymphoblastic leukemia, cardiomyopathy, ileal volvulus, and pancreatitis.
- Children with a misdiagnosis had similar amounts of stool on AXR as those who were not misdiagnosed.
- AXR was performed more frequently in those with a misdiagnosis (75% vs. 46%).
- Rectal examination was documented in only 9% of those with a diagnosis of constipation (low frequency rectal examination has been shown in other ED-based studies).
- Abdominal pain and tenderness were more common in those with a misdiagnosis.
Why I think this study is important:
While the authors point out that 1 in 200 children ultimately required a surgical or radiologic intervention within 7 days, I do not think that this error rate or diagnostic delay is particularly high. What is important is that this study reiterates the fact that AXRs are not useful for the diagnosis of constipation. The authors note “reviews have concluded that there is no evidence of an association between clinical symptoms of constipation and fecal loading on AXR.” Furthermore, AXRs may lead ER physicians to a cognitive diagnostic error.
Also, the misdiagnosis rate is much greater than reported in the study due to the definition adopted by the authors. The authors did not include treatable infectious diseases (e.g.. pneumonia, urinary tract infection) as well as a large number of other medical diagnosis. Other “incipient” disease processes may have been missed including inflammatory bowel disease and celiac disease.
The authors imply that using a more standard definition of constipation would be useful, namely the Iowa criteria which requires the presence of 2 of the following:
- ❤ stools/week,
- ≥1 episoded of fecal incontinence/week
- large stool palpable on rectal/abdominal examination
- passing large stool which obstructs toilet
- withholding posturing
- painful defecation
The authors reference a study which indicated that AXRs should be restricted to patients with high-yield clinical features: prior abdominal surgery, foreign body ingestion, abnormal bowel sounds, addominal distention, and peritoneal signs.
Bottomline: AXRs do not establish a diagnosis of constipation. Yet, after families have been told their child is constipated because of an AXR it is not easy to convince them that an AXR is about as useful as a pencil test for this diagnosis.
Related blog posts:
- Miralax Safety | gutsandgrowth
- Diagnostic tests hardly ever help patients poop | gutsandgrowth
- Updated Pediatric Expert Constipation Guidelines | gutsandgrowth
- AGA Constipation Guidelines | gutsandgrowth
- Data Supporting Miralax | gutsandgrowth
- Newsflash: constipation frequently causes … – gutsandgrowth