Pre-Test Probability And The Probe-To-Bone Test: What You Should Know

Ever wonder why there is such controversy over the probe-to-bone test? What could be simpler than inspecting a diabetic foot ulcer for bone with a sterile, blunt-ended metal probe? This is likely a test each podiatrist uses daily as it is a quick and cheap bedside exam. However, the ability to perform a simple test and the ability to interpret a simple test require a very different set of skills, and are by no means interchangeable. In fact, the interpretation depends heavily upon the population in which it is used.

Residents and students are often asked about the classic articles from Grayson and Lavery and their respective coworkers.^1,2 These two articles emphasize an important concept of any diagnostic test: pre-test probability that influences how clinicians use tests in clinical practice.

Grayson and colleagues looked at 75 hospitalized patients with 76 infected ulcerations.¹ Using the probe-to-bone test, the authors noted that 33 of 50 patients had a positive test subsequently confirmed by additional findings and only four of the 26 patients without additional findings of osteomyelitis had a positive test. This yielded a 66 percent sensitivity and 85 percent specificity corresponding to an 89 percent positive predictive value and 56 percent negative predictive value. In layman terms, this study showed that in hospitalized patients (already sick patients), if you probe to bone, you can be pretty certain osteomyelitis is present but if you don’t probe to bone, you’re still not sure.

Twelve years later, Lavery and coworkers conducted a similar study in an outpatient setting looking at 247 patients with ulcerations.² Of the 30 ulcerations with proven osteomyelitis via bone biopsy, 26 ulcers probed to bone but of the 217 ulcerations without osteomyelitis, 20 probed to bone (a false positive). This corresponds to 87 percent sensitivity, 91 percent specificity, a 57 percent positive predictive value and a 98 percent negative predictive value. This is nearly opposite of the conclusions of Grayson and colleagues. Essentially, the findings of Lavery and coworkers suggest that when an ulcer probes to bone, you still don’t know if the bone is infected (you’re only 57 percent sure). However, if you don’t probe to bone, you can be pretty certain (98 percent) there is no bone infection.

Throughout the last 10 years, multiple studies have also looked at the probe-to-bone test and found that in fact both studies are representative of how the probe to bone test works.^1-7 More recently, a systematic review has conveniently packaged this information in an easily digestible form.⁸ As the prevalence of osteomyelitis in the study increases (such as in sick inpatients), so does the ability to use the probe to bone test. However, as the prevalence of osteomyelitis decreases to low levels (such as in outpatients), one can no longer use the probe-to-bone test this way but can use it to exclude bone infection. This is both an incredibly complicated and simple phenomenon to comprehend. Looking at the mathematical reason why this occurs seems daunting but when looking at in a more practical application, this is just common sense.

Simplifying The Concept Of Pre-Test Probability

Consider, for one moment, the use of the metal detector in two different scenarios. In the first scenario, when the metal detector alarm goes off as it often does at the airport, it very seldom means someone with a weapon with intent to harm has passed through. In fact, individuals without an alarm are presumed to be safe and an alarm indicates further investigation. This is in part because there are very few individuals who are attempting to carry through weapons at the airport.

Now in the second scenario, consider the use of the same metal detector for the entrance of a known criminal hideout. Certainly, when the metal detector goes off, the suspicion of a person carrying a weapon with the intent to harm is much more likely because of the population one is using the metal detector on.

In a similar vein, hospitalized patients are more likely to have osteomyelitis. Likewise, when using the probe-to-bone test, the population of the patients you use it on matters.  

This is not just a principle exclusive for the probe-to-bone test. This applies to all diagnostic tests to some level. Understanding that pretest probability also affects studies on the erythrocyte sedimentation rate, radiographs and even magnetic resonance imaging (MRI) can help provide an accurate diagnosis in the midst of conflicting results, and provide insights into what the next step in care should be.

Kapoor and colleagues reported this phenomenon in a meta-analysis looking at MRI in the diagnosis of osteomyelitis.⁹ Clinicians frequently consider MRI to be the penultimate diagnostic test for osteomyelitis diagnosis (second to bone biopsy). However, if the pre-test probability is 10 percent, then a positive MRI for osteomyelitis increases the probability of actually having a bone infection to a mere 36.2 percent whereas a negative MRI will virtually exclude osteomyelitis by bringing the post-test probability to 1.3 percent.⁹ Conversely, when one uses MRI in a case in which the pretest probability is 75 percent, a positive MRI increases the post-test probability to 93.9 percent and a negative MRI results in a post-test probability of osteomyelitis of 26.5 percent.  

Once again, one can use MRI to rule out osteomyelitis when you believe it is not there and confirm osteomyelitis when you think it is present. Surely, this agrees with the age old saying, “treat the patient, not the lab/X-ray/etc.”

Applying The Pre-Test Probability

Using a pre-test probability implies that one already knows the prevalence of disease (in this case osteomyelitis) in a population. However, we evaluate patients individually and do not necessarily know the pre-test probability of our population. Instead, we extrapolate data from published studies and use it to help define our patient population in different settings such as the clinic versus hospital. A surrogate marker may be having a level of suspicion of the disease. The more the clinical scenario appears to assimilate to that of osteomyelitis through a history and physical, the more likely these findings indicate a higher prevalence of osteomyelitis in a group of all patients with these findings.

Additionally, the use of a multiples test may allow for a more closely approximated risk of osteomyelitis. Multiple authors have recommended this approach of combining clinical, radiographic and laboratory data.^5,10,11 Some have suggested the possibility that this combined diagnostic criteria may be more appropriate rather than any one test alone for osteomyelitis given that each test does not provide stand-alone proof of disease.¹¹ Diagnostic criteria may not only decrease the dependence on one test but may also reduce the ordering of unnecessary tests when the probability of osteomyelitis is already extremely likely or extremely unlikely. Additionally, one should also understand that in the face of conflicting evidence for and against the presence of osteomyelitis, a biopsy may be necessary as an unacceptable level of doubt may exist even with the use of a MRI.

Meyr and colleagues have noted that the once gold standard biopsy reputation has been tarnished by the uncomfortably low concordance between pathologists on what is deemed to be osteomyelitis.¹² However, the discussion of what is a biopsy and how biopsies are taken and interpreted is beyond the scope of this article.

In Conclusion

The appropriate use of a diagnostic tool, such as probe to bone, in relation to the prevalence of the disease is important for any clinical scenario. The difficulties of diagnosing osteomyelitis and the many different presentations that occur with this disease exemplify the need for considering pre-test probability prior to applying findings of a study to practice. Clinicians may apply this seemingly obscure principle of diagnostic testing much more broadly than its application to just the probe-to-bone test and osteomyelitis. Indeed, it may require one to reevaluate the application of every exam, lab test and radiographic finding to ensure appropriate use in the intended context.

Dr. Thorud is board-certified by the American Board of Podiatric Medicine and is an Associate of the American College of Foot and Ankle Surgeons.

Dr. Nguyen is an Associate of the American College of Foot and Ankle Surgeons. She is in private practice in Santa Cruz and Watsonville, Calif.

References

1. Grayson ML, Gibbons GW, Balogh K, et al. Probing to bone in infected pedal ulcers. A clinical sign of underlying osteomyelitis in diabetic patients. JAMA. 1995;273(9):721-723.
2. Lavery LA, Armstrong DG, Peters EJ, Lipsky BA. Probe-to-bone test for diagnosing diabetic foot osteomyelitis: reliable or relic? Diabetes Care. 2007;30(2):270-274.
3. Shone A, Burnside J, Chipchase S, et al. Probing the validity of the probe-to-bone test in the diagnosis of osteomyelitis of the foot in diabetes. Diabetes Care. 2006;29(4):945.
4. Morales Lozano R, Gonzalez Fernandez ML, Martinez Hernandez D, et al. Validating the probe-to-bone test and other tests for diagnosing chronic osteomyelitis in the diabetic foot. Diabetes Care. 2010;33(10):2140-2145.
5. Aragon-Sanchez J, Lipsky BA, Lazaro-Martinez JL. Diagnosing diabetic foot osteomyelitis: is the combination of probe-to-bone test and plain radiography sufficient for high-risk inpatients? Diabet Med. 2011;28(2):191-194.
6. Mutluoglu M, Uzun G, Sildiroglu O, et al. Performance of the probe-to-bone test in a population suspected of having osteomyelitis of the foot in diabetes. J Am Podiatr Med Assoc. 2012;102(5):369-373.
7. Zaiton F SA, Tarek TH, Tawfik AM, Hadhoud KM. Evaluation of diabetic foot osteomyelitis using probe to bone test and magnetic resonance imaging and their impact on surgical intervention. Egyptian J Radiol Nucl Med. 2014;45(3):795-802.
8. Lam K, van Asten SA, Nguyen T, et al. Diagnostic accuracy of probe to bone to detect osteomyelitis in the diabetic foot: a systematic review. Clin Infect Dis. 2016;63(7):944-948.
9. Kapoor A, Page S, Lavalley M, et al. Magnetic resonance imaging for diagnosing foot osteomyelitis: a meta-analysis. Arch Intern Med. 2007;167(2):125-132.
10. Fleischer AE, Didyk AA, Woods JB, et al. Combined clinical and laboratory testing improves diagnostic accuracy for osteomyelitis in the diabetic foot. J Foot Ankle Surg. 2009;48(1):39-46.
11. Berendt AR, Peters EJ, Bakker K, et al. Diabetic foot osteomyelitis: a progress report on diagnosis and a systematic review of treatment. Diabetes Metab Res Rev. 2008;24(Suppl 1):S145-161.
12. Meyr AJ, Singh S, Zhang X, et al. Statistical reliability of bone biopsy for the diagnosis of diabetic foot osteomyelitis. J Foot Ankle Surg. 2011;50(6):663-667.