Journal Logo

Research Article: Systematic Review and Meta-Analysis

Genetic polymorphisms in serine protease inhibitor Kazal-type 5 and risk of atopic dermatitis

A meta-analysis

Li, Yunling MD; Li, Yin MM; Li, Wei MM; Guo, Xiaoxuan MM; Zhou, Sha MM; Zheng, Huiwen MD

Editor(s): Saranathan., Maya

Author Information
doi: 10.1097/MD.0000000000021256
  • Open


Key points

  • The role of SPINK5 in atopic dermatitis (AD) risk was meta-analyzed.
  • SPINK5 Asn368Ser was significantly associated with AD risk.
  • Polymorphisms of Asp386Asn and Glu420Lys were not associated with AD risk.

1 Introduction

Atopic dermatitis (AD), also known as atopic eczema, is considered a chronic inflammatory skin disease. It results in dry, itchy, swollen, and red skin. As of 2018, the point prevalence of adult AD in the overall/treated populations was 4.9%/3.9% in the United States, 3.5%/2.6% in Canada, 4.4%/3.5% in the EU, and 2.1%/1.5% in Japan,[1] and the prevalence of AD continues to increase in developing countries.[2] Traditionally, AD is often associated with abnormalities in the skin barrier and immune system dysfunction, accompanied by high microbial colonization and a higher susceptibility to skin infection.[3,4] However, the pathogenesis of AD is not fully understood.

Serine protease inhibitor Kazal-type 5 (SPINK5) is a member of the gene family serine protease inhibitor Kazal-type cluster located on chromosome 5q32, which encode inhibitors of serine proteases. The encoded proteins are mainly distributed in the vaginal epithelium, thymus, vestibular gland, oral mucosa, tonsils, and parathyroid glands, which are mainly involved in the hydrolysis of human growth hormone and skin desquamation.[5] Various mutations in SPINK5 have been identified in patients with AD, and results were widely variable. For example, data from the study by Nishio et al. showed that five missense mutations, such as Asn368Ser, Asp386Asn, and Glu420Lys were associated with AD.[6] However, Jongepier and his colleagues demonstrated that SPINK5 was not associated with atopic phenotypes in individuals ascertained by a proband with asthma.[7] Thus, the association of AD with SPINK5 polymorphisms remains unclear, and results are not conclusive. Therefore, a meta-analysis would be needed to evaluate the role of SPINK5 polymorphisms and the risk of AD.

In this meta-analysis, previous studies associated with SPINK5 mutations (Asn368Ser, Asp386Asn and Glu420Lys) and AD risk were searched. An odds ratio (OR) with a 95% confidence interval (95% CI) was chosen as the effect size to explore the potential association between SPINK5 polymorphisms and AD risk.

2 Methods

The meta-analysis was performed following the guidelines provided by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P). Ethical approval was not necessary since this is a meta-analysis and no patients or animals involved.

2.1 Search strategy

Electronic English works of literature were searched from databases, including PubMed (, Embase (, and Cochrane library ( with a retrieval deadline of April 22, 2019, based on the predefined search strategy. The keywords and search terms used for all searches were “atopic dermatitis” or “AD” AND “SPINK5” OR “serine protease inhibitor kazal type 5” or “serine protease inhibitor Kazal-type 5” and “SNP” or “Single Nucleotide Polymorphism” or “polymorphism” or “genetic” or “variant”. Finally, in order to enroll more studies, articles of paper literature, and citations were manually screened.

2.2 Inclusion and exclusion criteria

The present meta-analysis would include the following studies: (1) the research design was a case-control study; (2) the participants in the case group were patients diagnosed with AD, and the participants in the control group were healthy people or hospitalized patients diagnosed without AD; (3) the association between SPINK5 polymorphism and AD was investigated; (4) data associated with the genotype and allele frequency of SPINK5 polymorphism (Asn368Ser, Asp386Asn and Glu420Lys) were provided.

Types of literature would be excluded if they were (1) studies with incomplete data, and statistical analysis could not be performed; (2) reviews, letters, and/or comments. For duplicated publications, only the study with the most complete data, most updated data, or higher Newcastle-Ottawa Scale (NOS) score could be included.

2.2.1 Data extraction and quality evaluation

The authors independently evaluated all relevant articles and extracted relevant data: the first author name, publication year, study region, diagnostic criteria of AD, the detection method of genotype, the same size in the case group and control group, gender, age, and outcomes in each group. Any discrepancies would be resolved by discussion. The genetic polymorphisms mainly included Asn368Ser, Asp386Asn, and Glu420Lys.

The NOS was used to assess the quality of enrolled studies,[8] which was recommended by the Agency for Healthcare Research and Quality (AHRQ) for quality assessment of each case-control study. For example, NOS scores of 0-3, 4-6, and 7-9 represented low, moderate, and high-quality studies, respectively.

2.3 Statistical analysis

Data analyses in this study were performed using the R software package 3.12. We initially assessed whether the genotype distribution in the control group was in accordance with Hardy Weinberg Equilibrium (HWE) by using the Chi-square test.[9] ORs with their 95% CI[10] of the allele model, co-dominant model, recessive model, and dominant model were calculated in order to assess the relationship between SPINK5 polymorphisms and AD. The heterogeneity was assessed by Dixon's Q-test[11] and I2 test. We defined that significant heterogeneity occurred if P < .05 or I2 > 50%, and then data would be pooled by the random effects model.[12] If P value >.05 or I2 < 50%, data would be pooled by the fixed effect model.[13] The publication bias was evaluated using the Egger's linear regression test[14] with P > .05, indicating no publication bias. Sensitivity analysis was performed by eliminating one study at each defined interval. The results are stable if the outcomes did not change.

3 Results

3.1 Study selection

Figure 1 shows the process of study selection in detail. Initially, a total of 120 potentially relevant papers were retrieved (PubMed: n = 42; Embase: n = 78; Cochrane Library: n = 0). Twenty-six duplicate articles were excluded by screening the titles. Next, 68 irrelevant articles were excluded after reading the title and abstract. For the remaining 26 publications, 20 (5 animal studies, 2 non-case and control studies, 4 meta-analyses/reviews, 1 duplicated population study, and 8 studies in which genotype data could not be obtained) articles were excluded by reviewing the full text. Finally, 6 articles met the inclusion criteria[7,15–19] and were included in the meta-analysis.

Figure 1
Figure 1:
Literature search and study selection.

3.2 Study characteristics and quality assessment

The included study characteristics are collected in Table 1. All 6 included articles are case-control studies and good quality studies with a NOS score ranging from 6 to 9. These studies were conducted in China, Japan, the United States, and Germany, and were published between 2003 and 2018. A total of 1968 participants were enrolled in this meta-analysis, including 914 patients with AD in the case group and 1054 participants in the control group. The diagnostic criteria of AD were mainly Hanifin and Rajka Criteria.[20] The genotype detection methods were polymerase chain reaction (PCR)-restriction fragment length polymorphism analysis (PCR-RFLP) and/or PCR amplification. Most articles did not report on the gender ration. Table 2 shows the SPINK5 gene polymorphisms [A1103G (Asn368Ser), G1156A (Asp386Asn), and G1258A (Glu420Lys)] of each included study. It is worth noting that the genotype distribution in the control group of one study [19] deviated from HWE.

Table 1
Table 1:
Characteristics of included studies.
Table 2
Table 2:
Gene distribution of included studies.

3.3 Meta-analysis for the association between SPINK5 polymorphism and AD

The associations between AD and the genetic polymorphisms of different genetic models for SPINK5 were evaluated in this study. The allele model (Asn368Ser: G vs A; Asp386Asn: A vs G; Glu420Lys: G vs A), co-dominant model (Asn368Ser: GG vs AA, GA vs AA; Asp386Asn: AA vs GG, AG vs GG; Glu420Lys: GA vs AA, GG vs AA), recessive model (Asn368Ser: GG vs AA+GA; Asp386Asn: AA vs GG+AG; Glu420Lys: GG vs AA+GA), and dominant model (Asn368Ser: GG+GA vs AA; Asp386Asn: AA+AG vs GG; Glu420Lys: GG+GA vs AA) for Asn368Ser, Asp386Asn, and Glu420Lys were evaluated.

The heterogeneity test results showed significant heterogeneity among the genetic models of GA vs AA, GG vs AA, GG vs AA+GA, and GG+GA vs AA for Glu420Lys (P < .05, I2 > 50%, Table 3). Therefore, data among individual studies were pooled using the random effects model. The fixed effects model was applied to other genetic models (Table 3).

Table 3
Table 3:
Meta-analysis of the association between genetic polymorphism of SPINK5 and AD.

The pooled estimates for Asn368Ser of the allele model (G vs A: OR = 1.2643, 95% CI = 1.0666–1.4987, P = .0069), co-dominant model (GG vs AA: OR = 1.6609, 95% CI = 1.1736–2.3505, P = .0042; GA vs AA: OR = 1.5448, 95% CI = 1.1263–2.1189, P = .0070), and dominant model (GG+GA vs AA: OR = 1.5700, 95% CI = 1.1656–2.1146, P = .0030) indicated significantly statistical differences, while the pooled estimates of the recessive model (GG vs AA+GA: OR = 1.0557, 95% CI = 0.8388–1.3287, P = .6441) were not significantly different (Fig. 2). Furthermore, since the OR value and its 95% CI were both greater than 1, the mutation of Asn368Ser in SPINK5 was determined to be a risk factor for AD. No statistically significant difference was found in the other genetic models for Asp386Asn (Fig. 3) and Glu420Lys (Fig. 4, all P > .05). These results demonstrated that the genetic polymorphism of Asn368Ser of SPINK5 was significantly related to AD morbidity and is a risk factor for AD.

Figure 2
Figure 2:
Meta-analysis of the association between genetic models of Asn368Ser and atopic dermatitis. The significant results were marked with “☆”.
Figure 3
Figure 3:
Meta-analysis of the association between genetic models of Asp386Asn and atopic dermatitis.
Figure 4
Figure 4:
Meta-analysis of the association between genetic models of Glu420Lys and atopic dermatitis.

3.4 Sensitivity analysis and publication bias

In the sensitivity analysis, the meta-analysis results of the genetic model GA vs AA of Asn368Ser, AA+AG vs GG of Asp386Asn, and AG vs GG of Asp386Asn were changed, while the other results were not changed. These results indicated the relative stability and reliability of the results. Additionally, Egger's test (P > .05) showed that publication bias among studies was not significant (Table 3).

4 Discussion

In our study, 6 articles were included in the meta-analysis. The statistically significant difference between patients with AD and healthy participants was calculated for the Asn368Ser polymorphism of the allele model, co-dominant model, and dominant model. However, no significant difference was found in the recessive model for Asn368Ser and other genetic models for Asp386Asn and Glu420Lys. Thus, our data suggest the SPINK5 Asn368Ser polymorphism may be a risk factor for AD.

It is well known that the skin acts as an essential barrier against pathogens and exogenous agents. Moreover, skin barrier dysfunctions are one of the major factors involved in AD development.[2] The gene SPINK5 is located on chromosome 5q31-32, which encodes the skin barrier protein lympho-epithelial Kazal-type-related inhibitor (also known as serine protease inhibitor Kazal-type 5).[21,22]SPINK5 in the epidermis is primarily expressed in the stratum granulosum, where it functions as a protease. Thus, it is important in the cornification of epithelial differentiation and exfoliation.[23,24] Previous evidence demonstrated that SPINK5 could prevent an influx of pathogens based on the formation of the cornified cell envelope.[5] Evidence from a study by Mocsai et al. showed that skin barrier functions were related to total immunoglobulin E (IgE) levels.[25] Furthermore, Tanei and his colleagues showed that allergic inflammation, mediated by the level of IgE, was crucial in the pathobiology of AD.[26] Recently, Hubiche et al. found an association between SPINK5 E420K polymorphisms and high IgE serum levels.[27] Moreover, several extensive studies support the role of SPINK5 in the development of AD.[16,18] Thus, the mechanism and the potential role of SPINK5 should be fully elucidated in future studies.

Notably, significant heterogeneity was calculated among data evaluating SPINK5 Glu420Lys polymorphism genotypes, including GA vs AA, GG vs AA, GG vs AA+GA, and GG+GA vs AA. Recently, it was determined that AD in several ethnic groups displayed variant mutation spots and rates between populations.[28] Moreover, skin barrier dysfunctions introduced by virulence factors could also induce allergic inflammation via innate and adaptive immunity.[29] Thus, limited background information of enrolled patients from a different ethnicity might be possible sources of heterogeneity. Although no significant difference was found in the recessive model for Asn368Ser and other genetic models for Asp386Asn and Glu420Lys in the meta-analysis, further clinical data would also be needed to verify the conclusion.

Furthermore, limitations of this meta-analysis should be noted. Firstly, the enrolled number of patients was small, and subgroup analysis could not be performed. Secondly, the genotyping method was different among enrolled studies, and the sensitivity ability of each method varied, which might lead to false-negatives. Third, though we found the association of SPINK5 Asn368Ser polymorphism and risk of AD, whether this SNP could influence gene expression of SPINK5 should be further investigated. Therefore, a study with higher NOS scores and larger sample size would be needed.

In conclusion, our study supports the role of SPINK5 Asn368Ser polymorphism as one of the risk factors for patients with AD. Future studies fully elucidating the pathogenic mechanisms involved in the disease are needed.

Author contributions

YLL and HWZ designed the study. YLL received the fund and was a major contributor in drafting the manuscript. HWZ revised the manuscript. YL, WL and XXG searched the references, reviewed the references and extracted the data. SZ performed quality evaluation and statistical analysis. All authors reviewed and approved the final version of the manuscript.


[1]. Barbarot S, Auziere S, Gadkari A, et al. Epidemiology of atopic dermatitis in adults: Results from an international survey. Allergy 2018;73:1284–93.
[2]. Lopez Carrera YI, Al Hammadi A, Huang YH, et al. Epidemiology,dDiagnosis, and treatment of atopic dermatitis in the developing countries of Asia, Africa, Latin America, and the Middle East: a review. Dermatol Ther (Heidelb) 2019;9:685–705.
[3]. Thomsen SF, Ulrik CS, Kyvik KO, et al. Importance of genetic factors in the etiology of atopic dermatitis: a twin study. Allergy Asthma Proc 2007;28:535–9.
[4]. Saurat JH, Hanifin JM. Understanding atopic dermatitis: pathophysiology and etiology. Conclusion. J Am Acad Dermatol 2001;45: (1 Suppl): S67–8.
[5]. Moffatt MF. SPINK5: a gene for atopic dermatitis and asthma. Clin Exp Allergy 2004;34:325–7.
[6]. Nishio Y, Noguchi E, Shibasaki M, et al. Association between polymorphisms in the SPINK5 gene and atopic dermatitis in the Japanese. Genes Immun 2003;4:515–7.
[7]. Jongepier H, Koppelman GH, Nolte IM, et al. Polymorphisms in SPINK5 are not associated with asthma in a Dutch population. J Allergy Clin Immunol 2005;115:486–92.
[8]. Stang A. Critical Evaluation of the Newcastle-Ottawa Scale for the Assessment of the Quality of Nonrandomized Studies in Meta-Analyses. Eur J Epidemiol 2010;25:603–5.
[9]. Schaid DJ, Jacobsen SJ. Blased Tests of Association: comparisons of allele frequencies when departing from Hardy-Weinberg proportions. Am J Epidemiol 1999;149:706–11.
[10]. Liu T, Xu QE, Zhang CH, et al. Occupational exposure to methylene chloride and risk of cancer: a meta-analysis. Cancer Causes Control 2013;24:2037–49.
[11]. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Ann Intern Med 1997;127:820–6.
[12]. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 2004;22:719–48.
[13]. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88.
[14]. Egger M, Smith GD, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34.
[15]. Folster-Holst R, Stoll M, Koch WA, et al. Lack of association of SPINK5 polymorphisms with nonsyndromic atopic dermatitis in the population of Northern Germany. Br J Dermatol 2005;152:1365–7.
[16]. Zhao LP, Di Z, Zhang L, et al. Association of SPINK5 gene polymorphisms with atopic dermatitis in Northeast China. J Eur Acad Dermatol Venereol 2012;26:572–7.
[17]. Kato A, Fukai K, Oiso N, et al. Association of SPINK5 gene polymorphisms with atopic dermatitis in the Japanese population. Br J Dermatol 2015;148:665–9.
[18]. Dezman K, Korosec P, Rupnik H, et al. SPINK5 is associated with early-onset and CHI3L1 with late-onset atopic dermatitis. Int J Immunogenet 2017;44:212–8.
[19]. Morizane S, Ouchida M, Sunagawa K, et al. Analysis of All 34 Exons of the SPINK5 Gene in Japanese Atopic Dermatitis Patients. Acta Med Okayama 2018;72:275–82.
[20]. J.H., G.R.. Diagnostic features of atopic dermatitis. Acta Derm Venereol (Stockh) 2000;80:430–4.
[21]. Chavanas S, Bodemer C, Rochat A, et al. Mutations in SPINK5, encoding a serine protease inhibitor, cause Netherton syndrome. Nat Genet 2000;25:141–2.
[22]. Sprecher E, Chavanas S, DiGiovanna JJ, et al. The spectrum of pathogenic mutations in SPINK5 in 19 families with Netherton syndrome: implications for mutation detection and first case of prenatal diagnosis. J Invest Dermatol 2001;117:179–87.
[23]. Chao SC, Richard G, Lee JY. Netherton syndrome: report of two Taiwanese siblings with staphylococcal scalded skin syndrome and mutation of SPINK5. Br J Dermatol 2005;152:159–65.
[24]. Norgett EE, Kelsell DP. SPINK5: both rare and common skin disease. Trends Mol Med 2002;8:7.
[25]. Mocsai G, Gaspar K, Dajnoki Z, et al. Investigation of skin barrier functions and allergic sensitization in patients with hyper-IgE syndrome. J Clin Immunol 2015;35:681–8.
[26]. Tanei R, Hasegawa Y, Sawabe M. Abundant immunoglobulin E-positive cells in skin lesions support an allergic etiology of atopic dermatitis in the elderly. J Eur Acad Dermatol Venereol 2013;27:952–60.
[27]. Hubiche T, Ged C, Benard A, et al. Analysis of SPINK 5, KLK 7 and FLG genotypes in a French atopic dermatitis cohort. Acta Derm Venereol 2007;87:499–505.
[28]. Palmer CN, Irvine AD, Terron-Kwiatkowski A, et al. Common loss-of-function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat Genet 2006;38:441–6.
[29]. Hanifin J, Saurat JH. Understanding atopic dermatitis: pathophysiology and etiology. Introduction J Am Acad Dermatol 2001;45: (1 Suppl): S1.

atopic dermatitis; meta-analysis; polymorphism; risk factor; SPINK5

Copyright © 2020 the Author(s). Published by Wolters Kluwer Health, Inc.