What a bunch of bollocks! John Thomas Sign is of no use in detecting pelvic fractures

You are working in a busy emergency unit that sees a barrage of trauma patients every weekend night. Mr Sithole is brought in following a pedestrian-vehicle accident by the provincial ambulance service. He is 33 years old, has a GCS 15/15 and haemodynamically stable. You notice that he has been immobilised on a rigid board including a collar and head blocks but that his left leg is free from the board but placed in a traction splint. The paramedic informs you that they suspect a mid-shaft femur fracture as there is a deformity, and bruising, along the medial aspect of the upper thigh and hip on the side of impact.

Following the initial clinical examination, which seems to confirm the presence of a femur fracture, you order an AP view of the pelvis along with the standard views of the femur and chest.

When the x-rays are handed to you it is clear that a femur fracture is indeed present in the left leg. However, the pelvic x-ray worries you. There certainly seems to be a disruption in the pelvic ring but given the positioning of the patient and his femur fracture you aren't sure if there is a real pelvic fracture or just an artefact. Perhaps there is a diagnostic sign that could assist you in making the diagnosis?

The John Thomas Sign

An incidental and sometimes humorous finding on an anteroposterior (AP) pelvic X-ray view is that of the John Thomas Sign (JTS). This sign is defined as positive when the penile shadow points towards the side of a hip or pelvic fracture in male patients.

To better define the role, significance and statistical validity of the JTS in predicting the presence of hip or pelvic fractures in the trauma setting, a systematic review and meta-analysis were performed by our group at the Department of Emergency Medicine, University of the Witwatersrand. The results appear in the Journal of Clinical Orthopaedics and Trauma and are available online now.

Hip and pelvic fractures are a common Emergency Department (ED) presentation that is associated with a 10%-13.4% mortality. A combination of clinical and radiological findings is required to confirm the diagnosis of pelvic and hip fractures. A 90-98% sensitivity of an Anterior-Posterior (AP) pelvis X-ray in diagnosing fractures has been reported, whereas this sensitivity is increased to 100% with either computed tomography (CT) and/or magnetic resonance imaging (MRI).

With this in mind, it's difficult to understand what the purpose of this diagnostic sign would be, as the available imaging studies perform very well. Nonetheless, perhaps there are circumstances where the sign could be useful.

Studies considered for inclusion met the following criteria: (i) the studies were clinical publications and (ii) studies reported either the sensitivity, specificity or provided detail with regards the positive versus negative sign findings. All published studies relating to the topic, including letters to the editor and conference papers, and non-English publications were eligible for inclusion. Single case reports were however excluded.

A positive JTS was regarded as the presence of a penile shadow (on AP pelvic x-ray) pointing towards the side of the hip or pelvic fracture, whereas a negative JTS was defined as the concomitant presence of a penile shadow pointing towards the midline or the opposite (contralateral) side of the fracture.

Data extraction and methodological quality evaluation

The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were used to ensure that the systematic review was robust. Nine articles were finally included in this review. All except one study was published in English.

Meta-Analysis

Meta-analysis, in the most general sense, is the application of statistical tools to pooled data from general very different studies, with the intention to determine the overall effect of an intervention, exposure or diagnostic tool. We use statistical tools, as opposed to arithmetic because we know that there is some error in both the estimate of the effect in the source studies and the estimate of the effect in our summary.

Of course, it's important to quantify this error as best as possible so that we can inform our clinical decision making about the uncertainty (FYI nothing is absolute in medicine even diagnostic tests, but more on this in another post).

Sensitivities and specificities were extracted from the relevant studies, where available. These values were used to calculate the Log Odds Ratio which gives us a composite measure of the performance of a test.

The meta-analysis was completed entirely in R using the Metafor package and some custom code. If you are interested in using the same approach I recommend these tutorials to get you going:

• AC Del Re - A practical tutorial on conducting meta-analysis in R
• J Rickert - R and Meta-Analysis
• JR Polanin - A review of meta-analysis packages in R
• D Quintana - Conducting a meta-analysis with R (Youtube)

The JTS was positive in 1089 out of 1439 patients with a fracture with a pooled sensitivity of 75.7% (95% CI, 73.4%-77.9%). Remember that AP Pelvic x-rays alone have a sensitivity of 90-98%.

Only six studies were included in the meta-analysis. Those that were not included did not have the necessary data in the published reports. Odds Ratios varied widely from -2.60 [95% CI: -3.4 – -1.8] to 1.66 [95% CI: 1.0 – 2.3]. A summary effect model was computed from these studies, yielding a log OR of -0.03 [95% CI: -1.3 – 1.2].

We use log Odds Ratio here because the natural logarithm transformation allows us to assume the data are normally distributed, but that's as far as I'm going on the nerdy stuff here.

Clinical Application

To put this into clinical perspective lets revisit our case:

Mr Sithole, being a 33-year-old male who was struck by a vehicle, has a +/-2% chance of having a pelvic fracture (Arregui-Dalmases et al 2014) based on epidemiological studies. You notice that his penis points distinctly to the left, the same side as the femur fracture and impact. You conclude that he has a positive John Thomas sign.

Does this mean he has a pelvic fracture? No, not quite.

The JTS has a log Odds Ratio of -0.03 which translates to an Odds Ratio of 0.97 (95% CI: 0.27-3.46). Now for simplicity (although we could use a calculator or an app) we use the Fagan nomogram to estimate the probability of Mr Sithole having a pelvic fracture. Page and Attia (2003) describe how this method works in an excellent BMJ article and also explain how to calculate the Relative Risk ratio (which is not the same thing as the Odds Ratio). The basic approach is to start at the baseline (aka pre-test) probability and then draw a straight line through the odds ratio to get the post-exposure or post-test probability. It's also useful to draw in the upper and lower estimates to get an idea of the range.

Although the figure is very busy, you should appreciate that the JTS has added no information to your assessment as the pre and post probabilities of a pelvic fracture are essentially the same!

Of course, we should have anticipated this since the initial estimate of the log Odds Ratio was very close to 0 AND the confidence interval crossed from negative to positive. In simple English, this means the test adds no information to the background rate of fractures and the confidence interval indicates that the test has been shown to demonstrate both a positive and a negative correlation with an actual fracture.

So balls to this test!

A quick note to keep in mind when applying diagnostic tests: The range of the confidence interval multiplies the pre-test probability. So if you have a large interval and a low pre test prob. then you won't be able to say much about the test results and whether they are clinically relevant. However, if you take the same test, and apply it to a patient with a large pre-test probability, your test result is likely to be more clinically relevant. So maximise pre-test probability by selecting appropriate tests and using clinical tools.

Headline image and case courtesy of Dr Gagandeep Singh, Radiopedia.

Source:

Louis Gerber, Abdullah E. Laher, Callistus O.A. Enyuma, Jared McDowall, Sunday J. Aigbodion, Sean Buchanan, Ahmed Adam, The ‘John Thomas’ sign and pelvic fractures—Fact or humorous myth?: A systematic review and meta-analysis, In Journal of Clinical Orthopaedics and Trauma, 2017, ISSN 0976-5662, https://doi.org/10.1016/j.jcot.2017.10.008.