Using the EHR to Find Missed Genetic Causes of Short Stature

Andrew N. Dauber, MD, MMSc


March 25, 2019

Editorial Collaboration

Medscape &

This transcript has been edited for clarity.

Short stature is one of the most common reasons that a child will be referred to a pediatric endocrinologist. As doctors, we often do extensive workup of these patients. Unfortunately, for the majority of those we see in the clinic, we never find an underlying diagnosis.

Today I'm going to tell you about two different studies on using electronic health records (EHRs) to help find patients with missed genetic causes of their short stature.

The first study[1] was focused on finding girls with Turner syndrome in whom it had not yet been detected through routine clinical evaluation. Among rare pediatric disorders, Turner syndrome is relatively common, occurring in approximately 1 in 2500 women. Women with this syndrome are missing one copy of the X chromosome, which can happen in all of the cells of the body or just some of the cells, which is called mosaic Turner syndrome.

Turner syndrome has many health implications, including effects on the heart and kidneys, frequent ear infections, learning issues, short stature, and problems with puberty and fertility in the future. So it really is important to make this diagnosis for these girls and young women.

Despite physicians' best intentions, there are reasons that somebody might miss a diagnosis.

We took a cohort of children who were found through a search of the EHR at Cincinnati Children's Hospital. We specifically looked at girls who were referred to the endocrine clinic with a height less than 2 standard deviations from the mean, so, towards the bottom of the regular growth chart. Their heights were also more than 1 standard deviation from where they should have been on the basis of their parents' heights. We hypothesized that that would identify a group of girls who had a high risk of having Turner syndrome.

One would think that, clinically, all of these girls would have had appropriate testing done for Turner syndrome, which is called a karyotype. But we found that a significant percentage of the girls had never been tested despite seeing a pediatric endocrinologist, who is often familiar with this condition.

We were able to get DNA from 37 of 70 or so girls who did not have a karyotype done and found that three of those girls had Turner syndrome. This is a remarkable finding. When we did the math, up to 6% of girls who were referred and met these criteria could be having a misdiagnosis of Turner syndrome.

Now, this was done at an excellent institution with a great group of pediatric endocrinologists who know a lot about Turner syndrome. But I think it goes to show you that despite physicians' best intentions, there are reasons that somebody might miss a diagnosis. By integrating these types of electronic searches that prompt us about clinical clues, we can find girls who have this important diagnosis that is otherwise being missed. It will be interesting to see how many young girls could be identified at an earlier age than through routine clinical practice if we generalize this to the entire US population.

The second study[2] was very similar in intent, but instead of focusing on Turner syndrome, we looked at other specific subgroups of short stature to see if we could identify rare genetic causes.

This study was a collaboration between Cincinnati Children's Hospital, Boston Children's Hospital, and Children's Hospital of Philadelphia, where we looked at patients with short stature but very high levels (above the 90th percentile) of insulin-like growth factor 1 (IGF-1), a growth factor that mediates the action of growth hormone.

This is very unusual, so we were able to recruit only 10 participants from all three institutions. Among the 10, we found a genetic cause in three participants. Again, none of these cases had been diagnosed clinically. Two of the three had mutations in the receptor for IGF-1 itself.

So again, the important lesson is that there are clinical clues that might evade a clinician's suspicion just from seeing the patient in clinic. By building in these searches of discrete data from the EHR, we're able to find these patients that would otherwise go clinically undiagnosed. I am sure that the integration of EHRs in genomics will only grow in the future.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.