Response to letter to the editor

We would like to thank Dr. Landel for his thoughtful Letter to the Editor. It is exciting to have colleagues as passionate about the Sharp Purser Test as our authorship team. Not only do we hope this review can guide clinicians, but we hope it can guide researchers to develop projects that can give clinicians more confidence in the reliability, validity, and safety of the Sharp Purer Test.

The data support our conclusion that inter-rater reliability was poor. The intra-rater reliability of the Sharp Purser Test was at best moderate, and at worst poor. After a robust search strategy of five databases from database inception to 2018, we only identified two studies that investigated the reliability of the Sharp Purser Test. The first study had four examiners perform the Sharp Purser Test on 11 children with Down’s syndrome and yielded poor to moderate intra-rater reliability scores (k = 0.67, 0.45, 0.29, 0.67), and inter-rater reliability ranged from 0.09 to 0.67 [1]. It is important to note the difference between intra and inter-rater reliability [2]. Intra-rater reliability is the agreement of the examiner at multiple points of administering the test, whereas inter-rater reliability is the agreement between multiple examiners. When appraising the inter-rater reliability of one examiner compared to another, we see mostly poor inter-rater reliability (k = 0.17, 0.09, 0.67, 0.09, 0.17, 0.45) [1]. In the second reliability study, the inter-rater reliability of the Sharp Purser Test in individuals with rheumatoid arthritis was poor (k = 0.20) [3]. In addition, a recent study was unable to calculate the reliability of the Sharp Purser Test in adults with neck pain due to 100% prevalence of negative findings [4].

If we assume that a test that is unreliable cannot possibly be valid [2] then we must first ask the question, why is the reliability of the Sharp Purser Test suboptimal? It could be the multiple ways that a positive test can occur: 1) reproduction of myelopathic signs with flexion and reduction in symptoms with the sharp purser maneuver or 2) signs of ligamentous laxity with a sliding motion of head posteriorly which may produce a clunk when the dens approximate with the posterior aspect of the atlas [5]. We believe the first condition is likely to have good reliability. However the second condition for a positive Sharp Purser Test is potentially why the reliability suffers so significantly which is more difficult for examiners to detect [6]. This idea was supported by Matthews, where in a five year follow up study to the original article [6] that focused on the Sharp Purser Test, concluded that signs of clunking and abnormal laxity were unreliable [7]. We know that the diagnostic utility of the Sharp Purser Test improves with more severe cases of instability [8], but this likely increases the chance of myelopathic symptoms being produced. Perhaps a paradigm shift is necessary in categorizing what constitutes a positive Sharp Purser Test in order to improve the reliability [9].

Our review concluded that the validity of the test was inconsistent due to the variability in diagnostic utility [10]. Dr. Landel argued that there was consistency amongst the data we presented, and he was correct! Of the studies where sensitivity and specificity could be measured, it did appear that there was consistency, with specificity being relatively high (SP: 0.77, 0.71, 0.98, 0.96) [3,6,8,11]. However, the inconsistency in validity was based on the totality of diagnostic utility yielded by our review [10]. There was inconsistency in the data with wide ranging sensitivity scores (SN: 0.40, 0.19, 0.44, 0.69), positive (+LR: 0.65, 1.73, 22, 17.25) and negative (-LR: 1.14, 0.77, 0.57, 0.32) likelihood ratios [3,6,8,11]. The diagnostic utility scores of the Sharp Purser Test need to be considered with our methodological quality assessment [10]. Matthews [6], Forester [3], and Stevens [11], did not know the results of the radiographs prior to using the Sharp Purser Test, but it is unclear if Uitvlugt did or did not [8]. Additionally, we were unable to perform a meta-analysis on the data available due to statistical heterogeneity [12,13]

We concluded that the Sharp Purser Test ‘has the potential to cause harm.’ Dr. Landel had concerns about this statement since no adverse events occurred in the 32 studies included in our review and would prefer this statement be modified to the following: we cannot ‘make a definitive conclusion about the relative safety of the test.’ This modification to our conclusion is acceptable. In our opinion, our conclusion of the Sharp Purser Tests ‘potential to cause harm’ was not disingenuous but rather to convey the point that no one has researched the safety of this test adequately. Therefore, we could not endorse the safety of the test, which is why we used the word ‘potential’. Interestingly, after our manuscript was published, we did hear from clinicians where complications occurred as a result of using the Sharp Purser Test.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).