a b s t r a c t

Objective: We aimed to evaluate the prognostic value of neutrophil-to-lymphocyte ratio in emergency de- partment (ED) patients with cutaneous Adverse drug reactions to identify the severe patients at an early stage. Methods: In this retrospective study, patients aged 18 and over who admitted to the ED of a university hospital with the diagnosis of cutaneous adverse drug reaction were included. For included patients, clinical findings and ED admission complete blood count results were recorded. The primary outcome was hospitalization and the secondary outcome was the type of drug reaction.

Results: A total of 135 patients were included in the study. The median age of patients was 50 (36-64) years. There was no significant difference between the patients hospitalized and discharged from the ED in terms of age and gender (p = 0.340 and p = 0.762, respectively). There was no significant difference between hospital- ized and discharged patients in terms of complete blood count parameters (p > 0.05, for all). The median NLR of hospitalized patients was significantly higher than that of patients discharged from the ED (6.13 vs. 3.69, p = 0.006). The median NLR of the patients with erythema multiform/Steven Johnson syndrome/toxic epider- mal necrosis was significantly higher than the NLR of the patients with maculopapular and fixed drug erup- tions (p = 0.022 and p = 0.015, respectively). The area under the curve value of NLR in predicting hospitalization was 0.640 (0.546-0.734). For 8.4 of NLR cutoff value, specificity was 83.9%.

Conclusion: NLR is a useful and simple prognostic parameter as an indicator of systemic inflammatory involve- ment in ED patients with cutaneous Adverse drug reactions. NLR is a useful parameter for deciding which patient will be admitted to the hospital in that patient group.

(C) 2021

1. Introduction

Adverse drug reactions are considered to be costly health problems for healthcare systems worldwide [1-3]. Although all organs can be af- fected by adverse drug reactions, the skin is the most frequently affected organ. Many drug reactions follow a benign course, but 2% of cutaneous adverse drug reactions are serious and life-threatening [4]. While the most common drug reactions with benign course are maculopapular or urticarial rashes, severe cutaneous adverse drug reactions are Acute generalized exanthematous pustulosis , drug reaction with eo- sinophilia and systemic symptoms (DRESS), Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN) [5,6]. Although the mecha- nism of a small number of drug reactions can be proven, they are mostly immune-mediated hypersensitivity reactions. The underlying patho- physiological mechanisms are very diverse [4,7].

* Corresponding author at: Necmettin Erbakan Universitesi, Meram Tip Fakultesi, Acil Tip AD 42080, Meram, Konya, Turkey.

E-mail address: zerdef@hotmail.com (Z.D. Dundar).

Due to its high morbidity and mortality potential, rapid diagnosis of drug reactions and early intervention are of great importance, but it is not always easy to identify patients at an early stage and make the cor- rect diagnosis [8-11]. It is known that various indicators such as Glas- gow prognostic score, systemic inflammatory response syndrome score showing systemic inflammation help to determine the prognosis of patients. Although there are SJS/TEN scoring systems, there are no op- timum biomarkers defined in the current literature to predict the sever- ity of cutaneous adverse drug reactions [12].

Neutrophil-to-lymphocyte ratio is a simple parameter that in- dicates the systemic inflammatory state of patients. The studies show that NLR is a good marker to predict prognosis and mortality in many diseases, including cancer, infectious and inflammatory diseases [13]. Although many studies showing the prognostic value of NLR in many dermatoses as an inflammatory marker are available in the current liter- ature, there is limited data on the prognostic value of NLR in patients with cutaneous adverse drug reactions [12]. Komatsu-Fujii et al. re- ported that NLR values were 18.5 in patients with DRESS and 4.5 in pa- tients with maculopapular eruption, but the authors did not report any

https://doi.org/10.1016/j.ajem.2021.05.029

0735-6757/(C) 2021

Parameters	Total (n = 135)	Hospitalized (n = 79)	Discharged from ED (n = 56)	p value
Age, median (Q1-Q3)	50.0 (36.0-64.0)	48.0 (34.0-64.0)	51.5 (41.0-64.0)	0.340
Gender, male, n (%) Comorbidity, n (%)a	63 (46.7)	36 (45.6)	27 (48.2)	0.762
No comorbidity	53 (39.3)	28 (35.4)	25 (44.6)	0.290
Hypertension	21 (15.6)	14 (17.7)	7 (12.5)
Diabetes mellitus	16 (11.9)	11 (13.9)	5 (8.9)
Rheumatological diseases	16 (11.9)	13 (16.5)	3 (5.4)
Asthma/COPD	12 (8.9)	7 (8.9)	5 (7.1)
Coronary artery disease	9 (6.7)	8 (10.1)	1 (1.8)
Triggering drug, n (%) Only one Drug history

Antibiotics	43 (31.9)	28 (35.4)	15 (26.8)	0.063
Analgesics	33 (24.4)	14 (17.7)	19 (33.9)
Antiepileptics/Antidepressants	9 (6.7)	8 (10.1)	1 (1.8)
Others	22 (16.3)	12 (15.2)	7 (12.5)
More than one drug history	19 (14.3)	14 (17.7)	8 (14.3)
Unknown drug	9 (6.7)	3 (3.8)	6 (10.7)

Involvement area, n (%)^b

Oral mucosa	39 (28.9)	31 (39.2)	8 (14.3)	0.002
Face	17 (12.6)	13 (16.5)	4 (7.1)	0.108
Trunk	97 (71.9)	59 (74.7)	38 (67.9)	0.385
Genitalia	15 (11.1)	12 (15.2)	3 (5.4)	0.073
Extremity	99 (73.3)	58 (73.4)	41 (73.2)	0.979
Drug reaction type, n (%) Group 1	18 (13.3)	14 (17.7)	4 (7.1)	0.270
Group 2	10 (7.4)	5 (6.3)	5 (8.9)
Group 3	42 (31.1)	27 (34.2)	15 (26.8)
Group 4	44 (32.6)	24 (30.4)	20 (35.7)
Group 5	15 (11.1)	6 (7.6)	9 (16.1)
Other	6 (4.4)	3 (5.4)	3 (5.4)

Group 1, erythema multiform/Steven Johnson syndrome/toxic epidermal necrosis; Group 2, acute generalized exanthematous pustulosis /acneiform/pustular; Group 3, anaphylaxis/ angioedema/urticarial; Group 4, maculopapular; Group 5, fix drug eruption; ED, emergency department; COPD, chronic obstructive pulmonary disease

^a Some patients have two or more comorbidities.

^b Since some of the patients have more than one area involvement, the total percentage exceeds 100%.

cutoff value to predict hospitalization in patients with cutaneous ad- verse drug reactions [12].

In this study, we aimed to evaluate the prognostic value of NLR in emergency department (ED) patients with cutaneous adverse drug re- actions to identify the severe patients at an early stage.

1. Methods

This retrospective study conducted in the Emergency Medicine De- partment of a university hospital. The study protocol was approved by the local ethical committee (decision number: 2021/2995).

Patients aged 18 and over who admitted to the ED and requested consultation from the Dermatology Department with the diagnosis of cutaneous adverse drug reaction between 2014 and 2020 were included in this study. The clinical and laboratory findings of the patients meeting the inclusion criteria were evaluated retrospectively (by reviewing elec- tronic case files). During the file review phase, patients with known he- matological diseases, patients received chemotherapy in the last week before emergency department admission, and patients who had miss- ing complete blood count were excluded from the study.

For included patients, age, gender, existing comorbid diseases, drug history thought to trigger the rash, dermatological examination find- ings, eruption distribution areas, type of drug reaction, hospitalization information were recorded. Again in the same patient group, the results of the first complete blood count requested in the ED were recorded. The NLRs of patients were calculated using the recorded neutrophil and lymphocyte counts.

The primary outcome of the study was hospitalization and the secondary outcome of the study was the type of drug reaction. In order to evaluate the primary outcome, the patients were divided into two groups: hospitalized and discharged from the ED. For the sec- ondary outcome, the patients were grouped based on the drug reaction

type: erythema multiform (EM)/SJS/TEN (Group 1), AGEP/acneiform/ pustular (Group 2), anaphylaxis/angioedema/urticarial (Group 3), maculopapular (Group 4), and fix drug eruption (Group 5). The hema- tological parameters were compared between the patient groups.

Data were analyzed using SPSS 20.0 software. Categorical variables were expressed as n (%), numerical variables conforming to the normal distribution as mean +- standard deviation, and numerical variables not conforming to the normal distribution as median (25% -75% quartiles). The difference of complete blood count parameters between the groups hospitalized and discharged from the ED was investigated using the Mann-Whitney U test or the independent samples t-test. The difference between the groups in terms of the drug reaction type of complete blood count parameters was investigated using the Kruskal Wallis test or ANOVA test. The power of NLR to predict hospitalization was ana- lyzed with ROC analysis and area under the curve (AUC) was evaluated. Optimum cutoff value for NLR was calculated by using Youden’s index. p < 0.05 is considered as statistically significant.

1. Results

A total of 135 patients were included in the study. The median age of patients was 50 (36-64) years and 63 (46.7%) of them were male. 79.3% of the patients had a history of taking a single drug, and the leading drugs that triggered the reaction in those patients taking single drug were antibiotics (31.9%), analgesics (24.4%) and antiepileptic/antide- pressants (6.7%). The drug history that triggered the reaction could not be detected in 6.7% of the patients. While 39.3% of the patients had no history of comorbidity, the most common comorbidities were hypertension (15.6%), diabetes mellitus (11.9%) and rheumatological diseases (11.9%). In dermatological examination findings, 71.9% of the patients had torso involvement, 73.3% of the extremity involvement and 28.9% of the oral mucosa involvement. The most common diagnoses

Comparison of the complete blood counts of hospitalized and discharged patient groups.

Parameters	Hospitalized	Discharged from	p value
	(n = 79)	ED (n = 56)
White blood cell, 103/uL	9.6 (6.9-14.3)	8.9 (6.4-12.4)	0.365
Neutrophil, 103/uL	7.6 (5.1-10.4)	6.6 (4.5-9.5)	0.134
Lymphocyte, 103/uL	1.1 (0.7-2.1)	1.4 (1.0-2.6)	0.062
Monocyte, 103/uL	0.5 (0.2-0.8)	0.6 (0.3-0.8)	0.132
Eosinophil, 103/uL	0.06 (0.00-0.30)	0.10 (0.01-0.29)	0.534
Basophil, 103/uL	0.00 (0.00-0.04)	0.01 (0.00-0.08)	0.424
NLR	6.13 (3.40-11.49)	3.69 (2.46-7.24)	0.006

NLR, neutrophil-to-lymphocyte ratio; ED, emergency department.

were maculopapular (32.6%), anaphylaxis/angioedema/urticarial (31.1%) and EM/SJS/TEN (13.3%)0.79 (58.5%) of 135 patients were hos- pitalized. When patients hospitalized and discharged from the ED were compared, there was no significant difference between the groups in terms of age and gender (p = 0.340 and p = 0.762, respectively). Oral mucosal involvement was significantly higher in hospitalized patients than patients discharged from the ED (39.2% vs. 14.3%, p = 0.002). There was no significant difference in drug reaction type between hos- pitalized and discharged patients (p = 0.270). Of those 79 hospitalized patients, 37 (46.8%) patients stayed in hospital for less than 5 days. The demographic characteristics and clinical findings of the patients hospi- talized and discharged from ED are shown in Table 1.

The comparison of the complete blood count findings of the patient groups hospitalized and discharged from the ED is shown in Table 2. There was no significant difference between hospitalized and discharged patients in terms of white blood cell count (p = 0.365), neu- trophil (p = 0.134), lymphocyte (p = 0.062), monocyte (p = 0.132), eosinophil (p = 0.534) and basophil (p = 0.424). The median NLR of hospitalized patients was significantly higher than the median NLR of those discharged from the ED (6.13 vs. 3.69, p = 0.006) (Fig. 1). There was no difference between patients hospitalized for <5 days and pa- tients hospitalized for >=5 days in terms of NLR (6.23 (3.17-11.61) vs.

5.73 (3.53-11.53), respectively, p = 0.795). The AUC value of NLR in

Fig. 2. Receiver operating characteristic (ROC) curve of neutrophil-to-lymphocyte ratio in terms of hospitalization. Area under the curve is 0.640 (95% confidence interval 0.546-0.734). For 8.4 of NLR cutoff value, sensitivity is 40.5% and specificity is 83.9%.

predicting hospitalization was 0.640 (0.546-0.734) (p = 0.006) (Fig. 2). The optimum cutoff value of NLR was determined as 8.4 to pre- dict hospitalization. For 8.4 of NLR cutoff value, sensitivity was 40.5% and specificity was 83.9%.

The comparison of the complete blood count findings of the patient

groups according to the drug reaction type is shown in Table 3. White blood cell and neutrophil levels of the patients with AGEP were signif- icantly higher than the patient groups with other drug reaction types

Fig. 1. Comparison of the neutrophil-to-lymphocyte ratios of hospitalized and discharged patient groups. ED, emergency department.

Table 3

Comparison of the complete blood counts of patient groups according to drug reaction type.

Parameters	Group 1 (n = 18)	Group 2 (n = 10)	Group 3 (n = 42)	Group 4 (n = 44)	Group 5 (n = 15)
White blood cell, 103/uL	9.4 (6.8-11.6)a	13.7 (10.3-21.5)a	9.9 (6.9-13.8)a	7.9 (5.0-10.9)a	9.7 (7.2-11.7)a
Neutrophil, 103/uL	7.4 (6.2-9.9)	10.6 (7.7-18.3)b	7.8 (5.3-10.2)b	6.1 (3.6-8.4)b	6.2 (5.1-9.4)
Lymphocyte, 103/uL	0.9 (0.6-1.3)c	2.4 (1.6-2.8)c	1.5 (0.8-2.6)c	1.1 (0.7-1.9)	1.6 (0.9-3.0)c
Monocyte, 103/uL	0.4 (0.2-0.8)	0.9 (0.7-1.1)	0.5 (0.2-0.8)	0.5 (0.2-0.8)	0.6 (0.4-0.8)
Eosinophil, 103/uL	0.2 (0.0-0.3)	0.1 (0.0-0.2)	0.0 (0.0-0.1)	0.1 (0.0-0.4)	0.2 (0.0-0.4)
Basophil, 103/uL	0.0 (0.0-0.1)	0.0 (0.0-0.1)	0.0 (0.0-0.0)	0.0 (0.0-0.1)	0.0 (0.0-0.1)
NLR	8.3 (4.8-14.8)d	4.2 (2.9-11.5)	5.4 (2.9-10.1)d	4.2 (3.1-8.2)d	3.3 (2.7-5.9)

Group 1, erythema multiform/Steven Johnson syndrome/toxic epidermal necrosis; Group 2, acute generalized exanthematous pustulosis /acneiform/pustular; Group 3, anaphylaxis/an- gioedema/urticarial; Group 4, maculopapular; Group 5, fix drug eruption; NLR, neutrophil-to-lymphocyte ratio.

^a p values are 0.013 (Group 1 – Group 2), 0.014 (Group 2 – Group 3), 0.002 (Group 2 – Group 4), 0.008 (Group 2 – Group 5).

^b p values are 0.043 (Group 2 – Group 3), 0.002 (Group 2 – Group 4), 0.010 (Group 2 – Group 5).

^c p values are 0.001 (Group 1 – Group 2), 0.017 (Group 1 – Group 3), 0.024 (Group 1 – Group 5).

^d p values are 0.015 (Group 1 – Group 3), 0.022 (Group 1 – Group 4).

(p = 0.007 and p = 0.010, respectively). The lymphocyte values of the patients with EM/SJS/TEN were lower than the patient groups with other drug reaction types (p = 0.007). The median NLR of the patients with the EM/SJS/TEN was significantly higher than the NLR of the pa- tients with maculopapular and fixed drug eruptions (p = 0.022 and p = 0.015, respectively), but no significant difference was found be- tween the NLRs of patients with AGEP and anaphylaxis/angioedema/ urticarial (p = 0.195 and p = 0.095, respectively) (Fig. 3). There was no significant difference between triggering drug groups in terms of NLR (p = 0.621).

1. Discussion

In this retrospective observational study, we investigated the prog- nostic value of NLR in patients with cutaneous adverse drug reactions in the ED. There was no significant difference between hospitalized and discharged patients in terms of parameters involved in complete blood count; however we found that hospitalized patients had signifi- cantly higher NLR levels than those discharged from the ED. We also found that patients with EM/SJS/TEN, which are considered to be severe

cutaneous adverse drug reactions, were higher than patients with maculopapular and fixed drug eruptions.

Cutaneous adverse drug reactions are Rare conditions in EDs how-

ever early recognition of serious and life-threatening reactions is impor- tant [14]. Drugs that frequently trigger cutaneous adverse drug reactions are analgesics, antibiotics and antiepileptics [4]. In our study, similar to the literature, we found that the same three drug groups were responsible for the adverse drug reactions and antibiotics were the leading drugs in our study group.

Although the diagnosis of drug reactions is generally made by history and physical examination, it is important to determine the extent of sys- temic involvement in ED patients with cutaneous adverse drug reaction. Our study results showed that patients with more severe cutaneous ad- verse drug reactions had an increased NLR. Complete blood count is a sim- ple, rapidly resulting and easily performed test in any ED. There are many studies in the literature reporting that NLR is an indicator of systemic in- flammation in different diseases [15-19]. However, there are a limited number of studies showing the prognostic value of NLR in cutaneous ad- verse drug reactions (12,20). Komatsu-Fujii et al. reported that NLR of DRESS patients was higher than that of patients with maculopapular ex- anthema [12]. SJS/TEN patients were not included in their study and

Fig. 3. Comparison of the neutrophil-to-lymphocyte ratios of patient groups according to drug reaction type. Group 1, erythema multiform/Steven Johnson syndrome/toxic epidermal necrosis; Group 2, acute generalized exanthematous pustulosis /acneiform/pustular; Group 3, anaphylaxis/angioedema/urticarial; Group 4, maculopapular; Group 5, fix drug eruption.

therefore it is not possible to directly compare our results with their re- sults. However, the fact that NLR of patients with DRESS/DIHS in the study of Komatsu-Fujii et al. and that of patients with EM/SJS/TEN patients in our study were higher than those of patients with more benign diagno- ses supports the idea that the NLR ratio can be used as a systemic involve- ment indicator. Komatsu-Fujii et al. reported that the NLR level of patients with maculopapular exanthema was 4.5 [12] and we found that the NLR level was 4.2 in the same patient group. Our results are consistent with the results of that study.

Branicka et al. evaluated the eosinophil/neutrophil/platelet- lymphocyte ratios retrospectively in hypersensitivity reactions due to nonsteroidal anti-inflammatory drugs (NSAID) of 97 cases and they found that the NLR was insufficient to differentiate the groups of respi- ratory disease, cutaneous disease and urticaria/angioedema induced by NSAIDs [20]. However, in the same study, the authors reported that the evaluation of NLR and CRP together is useful in showing neutrophilic in- flammation [20]. There are various pathophysiological mechanisms for cutaneous adverse drug reactions [7] and our study results suggest that, regardless of the underlying mechanism, evaluating the severity of systemic inflammation with NLR is useful in predicting prognosis of patient with cutaneous adverse drug reaction.

On the other hand, differentiating patients who will be discharged from the ED and hospitalized is an important issue in the ED management of patients. Our study results showed that the NLR ratio is useful in show- ing the severity of systemic involvement in patients admitted to the ED with cutaneous drug reaction in addition to clinical findings and reveals that NLR is a parameter that can guide patients who need hospitalization. In addition, in our study, none of the white blood cell subgroups in those patients distinguish between hospitalized and discharged patients.

1. Conclusion

In conclusion, the NLR is a useful and simple prognostic parameter as an indicator of systemic inflammatory involvement in patients admitted to the ED with cutaneous adverse drug reactions. NLR is a useful param- eter for deciding which patient will be admitted to the hospital in that patient group.

Declaration of Competing Interest

The authors have no commercial associations or sources of support that might pose a conflict of interest.

References

1. Diaz L, Ciurea AM. Cutaneous and systemic adverse reactions to antibiotics. Dermatol Ther. 2012;25:12-22 2.
2. Carneiro SC, Azevedo-e-Silva MC, Ramos-e-Silva M. Drug eruptions in the elderly. Clin Dermatol. 2011;29:43-8.
3. Ahmed AM, Pritchard S, Reichenberg J. A review of cutaneous drug eruptions. Clin Geriatr Med. 2013;29:527-45.
4. Hoetzenecker W, Nageli M, Mehra ET, et al. Adverse cutaneous drug eruptions: cur- rent understanding. Semin Immunopathol. 2016;38(1):75-86.
5. Bircher AJ, Scherer K. Delayed cutaneous manifestations of drug hypersensitivity. Med Clin North Am. 2010;94(4):711-25.
6. French LE, Prins C. Erythema multiforme, Stevens-Johnson syndrome and toxic epi- dermal necrolysis. In: Bolognia JL, Jorrizo JL, Schaffer JV, editors. Dermatology. , 3rd ed.New York: Elsevier; 2013. p. 319-34.
7. Temiz SA, Ozer I, Ataseven A. Dermatologic emergencies. Selcuk Med J. 2020;36(2): 157-67.
8. Schneck J, Fagot JP, Sekula P, et al. Effects of treatments on the mortality of Stevens-Johnson syndrome and toxic epidermal necrolysis: a retrospective study on patients included in the prospective EuroSCAR Study. J Am Acad Dermatol. 2008;58:33e40.2.
9. Shiohara T, Iijima M, Ikezawa Z, et al. The diagnosis of a DRESS syndrome has been sufficiently established on the basis of typical clinical features and viral reactivations. Br J Dermatol. 2007;156:1083e4.3.
10. Mockenhaupt M. Stevens-Johnson syndrome and toxic epidermal necrolysis: clinical patterns, diagnostic considerations, etiology, and therapeutic management. Semin Cutan Med Surg. 2014;33:10e6.
11. Chu CY. Drug eruptions: great imitators. Clin Dermatol. 2020 Mar-Apr;38(2): 193-207. https://doi.org/10.1016/j.clindermatol.2019.10.005.
12. Komatsu-Fujii T, Chinuki Y, Niihara H, et al. The thymus and activation-regulated chemokine (TARC) level in serum at an early stage of a drug eruption is a prognos- tic biomarker of severity of systemic inflammation. Allergol Int. 2018;67(1):90-5.
13. Forget P, Khalifa C. DefourJ, et al. what is the normal value of the neutrophil-to- lymphocyte ratio? BMC Res Notes. 2017;10:12.
14. Mustafa SS, Ostrov D, Yerly D. Severe cutaneous adverse drug reactions: pre- sentation, risk factors, and management. Curr Allergy Asthma Rep. 2018, 24; 18(4):26.
15. Zhan H, Ma JY, Jian QC. Prognostic significance of pretreatment neutrophil-to- lymphocyte ratio in melanoma patients: a meta-analysis. Clin Chim Acta. 2018; 484:136-40.
16. Qin B, Ma N, Tang Q, Wei T, Yang M, Fu H, et al. Neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) were useful markers in assessment of inflammatory response and disease activity in SLE patients. Mod Rheumatol. 2016;26(3):372-6.
17. Galadari AA, Sun LD, Fernandez AP. Utility of the neutrophil/lymphocyte ratio for predicting malignancy risk and prognosis in patients with dermatomyositis. Clin Exp Dermatol. 2020 Jun;45(4):471-3.
18. Ozturk C, Balta S, Balta I, Celik T. neutrophil-lymphocyte ratio in Behcet disease. Angiology. 2015 Aug;66(7):695.
19. Coban AM, Aksoy E, Duman D, Ozmen I, Yildirum E, Gungor S, et al. Does eosino- philia and neutrophil to lymphocyte ratio affect hospital re-admission in cases of COPD exacerbation? Tuberk Toraks. 2017 Dec;65(4):282-90.
20. Branicka O, Rogala B, Gluck J. Eosinophil/neutrophil/platelet-to-lymphocyte ratios in various types of immediate hypersensitivity to NSAIDs: a preliminary study. Int Arch Allergy Immunol. 2020;181(10):774-82.