status epilepticus in re”>Erector spinae and serratus anterior blo”>NACA score and c”>tube thoracostomy device reduces i”>American Journal of Emergency Medicine 38 (2020) 1679-1694

Correspondence

Contents lists available at ScienceDirect

American Journal of Emergency Medicine

journal homepage:

Serum pentraxin-3 levels in acute cholecystitis: Helpful in emergency decision-making?

Dear Editor,

I have read the article “Relationship between serum Pentraxin 3 and pro-adrenomedullin levels with acute cholecystitis” by Algin et al. [1] with interest and would like to discuss the results of this study further with a statistical point of view. The authors concluded that serum levels of the two proposed markers, pentraxin-3 (PTX3) and pro- adrenomedullin (pro-ADM), may predict gallbladder perforation in

acute cholecystitis. For the cut-off values of 4.8 ng/mL for PTX3 and 97 nmol/mL for pro-ADM, the sensitivity-specificity of the tests were 75-85% and 100-95%, respectively. Ideally, a simple diagnostic/prog- nostic test should have zero false positive and False negative rates lead- ing to a specificity, positive predictive value, sensitivity, and negative predictive value of 100%. In the case of a continuous output such as the serum concentration of a biochemical marker, no intersection be- tween the positive and negative histograms for a situation would lead to a perfect test with 100% accuracy (Fig. 1). Unfortunately, almost no laboratory test based on the measurement of a serum marker has 100% diagnostic/Prognostic accuracy in clinical medicine due to some degree of intersection between the positive and negative results. It is the degree of this intersection that determines the performance of

Fig. 1. An ideal test with a continuous output divided into quantiles for differentiating Situation A from Situation B. No intersection between the positive and negative histograms for a situation leads to 100% accuracy if any cut-off in the no intersection area (grey zone) is applied. PPV = Positive Predictive Value, NPV = Negative Predictive Value.

0735-6757/(C) 2019

1680 Correspondence / American Journal of Emergency Medicine 38 (2020) 1679-1694

Fig. 2. The degree of the intersection between the positive and negative histograms (grey zone) determines the general performance of a test with a continuous output for a situation. The cut-off values above the upper dashed line give 100% specificity and positive predictive value (PPV), whereas the cut-off values below the lower dashed line give 100% sensitivity and negative predictive value (NPV) for predicting Situation A.

such a test. Different cut-off values within the intersection yield differ- ent sensitivity, specificity, positive and negative predictive values, and accuracy levels (Fig. 2). According to the expectation from a proposed test, a cut-off of higher sensitivity, specificity, or accuracy may be set. In general, the cut-off for a screening test is set at a relatively high sen- sitivity level at the expense of specificity, whereas more specific and ac- curate cut-offs are considered for diagnostic/prognostic tests [2]. An accuracy-based determination of a cut-off value for pro-ADM for the di- agnosis of gallbladder perforation in acute cholecystitis, hence the prog- nosis of cholecystitis, in the present study [1] is a good example for such an approach. Note that the distribution of serum pro-ADM in cases with and without perforation resembles the situation in Fig. 2. Serum PTX3 levels, on the other hand, follows a quite different distribution as com- pared to pro-ADM in cases with and without perforation (Fig. 3). Serum PTX3 follows a logarithmic distribution both in the healthy pop- ulation and patient groups with various diseases [3-7]. It has a wide normal range and may increase up to 1000 to 10,000 times in case of disease [3-6] unlike pro-ADM. As depicted in Fig. 3, the distribution of serum PTX3 in cholecystitis patients with no perforation intersects to- tally with that of patients with perforation in the present study [1]. Moreover, the median serum PTX3 levels of patients with and without perforation are too close to each other although there exists a significant difference between their distributions. Note that decreasing the cut-off to any lowest value would never give 100% sensitivity in terms of

prediction of perforation. Some cases with perforation may even have lower serum PTX3 levels than that of cases with no perforation (Fig. 3). So, serum PTX3 is far from being a reliably sensitive test for predicting perforation in acute cholecystitis. Increasing the cut-off above the upper limit of the cases with no perforation, 8.36 ng/mL, would still give 100% specificity at the expense of sensitivity (Fig. 3). But reproduction of this study with a higher number patients (i.e. with lower type I and type II error rates) would probably lead to an im- pairment in the specificity as compared to that of the present study since a considerable number of cases with serum PTX3 levels higher than 8.36 ng/mL would be expected to appear. Note that the minimum, median, and maximum values of serum PTX3 were 0.01, 0.73, and

60.76 ng/mL, respectively, in the control group of our previously pub- lished study and 14% of these healthy people had serum PTX3 levels higher than 8.36 ng/mL [3,4]. This rate may predictably be higher in dis- ease groups with Inflammatory conditions [7-9]. Another important point to be discussed is the lack of information regarding comorbid dis- eases such as diabetes, coronary artery disease, chronic obstructive pul- monary disease, and chronic kidney disease in the present study. As an example, diabetes is related to poor prognosis and perforation in acute cholecystitis [10]. It is also known that patients with diabetes have higher serum PTX3 levels compared to healthy controls [11,12]. The lack of control for confounders like diabetes complicates the interpreta- tion of the study results.

Correspondence / American Journal of Emergency Medicine 38 (2020) 1679-1694 1681

Fig. 3. The possible distributions of serum pentraxin 3 levels according to the presence of perforation were represented here by using the minimum, median, and maximum values reported in the study [1] with an assumption of the similarity of the distributions to those reported in previous studies regarding serum pentraxin 3 levels. Note that the distribution of serum pentraxin 3 in cholecystitis patients with no perforation intersects totally with that of patients with perforation. Decreasing the cut-off to any lowest value would never give 100% sensitivity in terms of prediction of perforation. Some cases with perforation have lower serum pentraxin 3 levels than that of cases with no perforation. Increasing the cut-off above the upper limit of the cases with no perforation (the upper dashed line which corresponds to 8.36 ng/mL) still gives 100% specificity and positive predictive value (PPV) at the expense of sensitivity.

Acknowledgments

The author thanks Prof. Selcuk Dagdelen for the introduction to the statistical basics in good medical practice sessions in medical school years.

Declaration of Competing of Interest

None.

Funding

None.

Ufuk Ilgen?

Department of Rheumatology, Trakya University Medical School, Edirne, Turkey E-mail address: [email protected].

16 March 2020

https://doi.org/10.1016/j.ajem.2019.158501

References

1. Algin A, Gulacti U, Inan I, Erdogan MO, Colak S, Sariaydin M. Relationship between serum Pentraxin 3 and pro-adrenomedullin levels with acute cholecystitis. Am J Emerg Med 2019;37:1268-72. https://doi.org/10.1016/j.ajem.2018.09.024.
2. Trevethan R. Sensitivity, specificity, and predictive values: foundations, pliabilities, and pitfalls in research and practice. Front Public Health 2017;5:307. https://doi. org/10.3389/fpubh.2017.00307.
3. Ilgen U, Yayla ME, Duzgun N. Low serum fibroblast growth factor 2 levels not ac- companied by increased serum pentraxin 3 levels in patients with systemic scle- rosis. Clin Rheumatol 2017;36:367-72. https://doi.org/10.1007/s10067-016- 3483-7.
4. Ilgen U, Yayla ME, Gokalp C, Emmungil H, Duzgun N. Serum pentraxin-3 follows a logarithmic distribution particularly at low expected levels. Int Wound J 2019. In press. https://doi.org/10.1111/iwj.13184.
5. Yamasaki K, Kurimura M, Kasai T, Sagara M, Kodama T, Inoue K. Determination of physiological plasma pentraxin 3 (PTX3) levels in healthy populations. Clin Chem Lab Med 2009;47:471-7. https://doi.org/10.1515/CCLM.2009.110.
6. Kontny F, Andersen T, Ueland T, Akerblom A, Lakic TG, Michelsen AE, et al. Pen- traxin-3 vs C-reactive protein and other Prognostic biomarkers in acute coro- nary syndrome: a substudy of the Platelet Inhibition and patients outcomes (PLATO) trial. Eur Heart J Acute Cardiovasc Care 2019. https://doi.org/10.1177/ 2048872619846334 In press.
7. Hamed S, Behnes M, Pauly D, Lepiorz D, Barre M, Becher T, et al. Diagnostic value of Pen- traxin-3 in patients with sepsis and septic shock in accordance with latest sepsis-3 defini- tions. BMC Infect Dis 2017;17:554. https://doi.org/10.1186/s12879-017-2606-3.
8. Ozer Balin S, Sagmak Tartar A, Ugur K, Kilinc F, Telo S, Bal A, et al. Pentraxin-3: a new parameter in predicting the severity of diabetic foot infection? Int Wound J 2019; 16:659-64. https://doi.org/10.1111/iwj.13075.
9. Oztan MO, Aksoy Gokmen A, Ozdemir T, Muderris T, Kaya S, Koyluoglu G. Pentraxin-3: a strong novel biochemical marker for appendicitis in children. Am J Emergency Med 2019. https://doi.org/10.1016/j.ajem.2019.01.010.
10. Karamanos E, Sivrikoz E, Beale E, Chan L, Inaba K, Demetriades D. Effect of diabe- tes on outcomes in patients undergoing emergent cholecystectomy for acute cholecystitis. World J Surg 2013;37:2257-64. https://doi.org/10.1007/s00268- 013-2086-6.
11. Wang R, Zhang J, Hu W. Association of serum pentraxin 3 concentrations with dia- betic nephropathy. J Investig Med 2016;64:1124-7. https://doi.org/10.1136/jim- 2016-000082.
12. Erdenen F, Gungel H, Altinoglu H, Sak D, Muderrisoglu C, Koro A, et al. Associa- tion of plasma pentraxin-3 levels with retinopathy and systemic factors in dia- betic patients. Metab Syndr Relat Disord 2018;16:358-65. https://doi.org/10. 1089/met.2018.0023.