G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
Journal of Clinical Virology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Review
HPV and lung cancer risk: A meta-analysis Kan Zhai a,∗,1 , Jie Ding a,1 , Huan-Zhong Shi a,b,∗ a b
Medical Research Center, Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China Department of Respiratory and Critical Care Medicine, Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China
a r t i c l e
i n f o
Article history: Received 30 April 2014 Received in revised form 21 September 2014 Accepted 22 September 2014 Keywords: HPV Lung cancer Risk Meta-analysis
a b s t r a c t The potential causal association between human papillomavirus (HPV) and lung cancer (LC) remains controversial. We performed this meta-analysis to evaluate whether HPV infection in lung tissue is associated with LC compared with non-cancer controls. We also quantified this association in different LC subtypes. MEDLINE, EMBASE and Web of Science were searched through March 2014, using the search terms “lung cancer”, “human papillomavirus”, “HPV” and their combinations. Association was tested using odds ratio (OR) with 95% confidence intervals (95% CI). Heterogeneity was assessed using Q and I2 statistic. Finally, nine studies, for a total of 1094 LCs and 484 non-cancer controls, were identified as eligible publications. The pooled results showed that HPV infection was associated with LC (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001). Similar results were also observed in HPV16 and/or HPV18 (HPV16/18) infection analyses (OR = 6.02, 95% CI: 3.22–11.28, P < 0.001). HPV16/18 was significantly associated with lung squamous cell carcinoma (SCC) (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001), while the pooled OR was 3.69 in lung adenocarcinoma (95% CI: 0.99–13.71, P = 0.052). Our results suggest that lung tissue with HPV infection has a strong association with LC, and especially, HPV16/18 infection significantly increases SCC risk, which indicates a potential pathogenesis link between HPV and LC. © 2014 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Study design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. 3.2. Study selection and inclusion criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Eligible studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. HPV infection and LC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Sensitivity analysis and publication bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Authors’ contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Competing interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ethical approval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
∗ Corresponding authors at: Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China. Tel.: +86 10 85231624. E-mail addresses: kan [email protected] (K. Zhai), [email protected] (J. Ding), [email protected] (H.-Z. Shi). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.jcv.2014.09.014 1386-6532/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
2
1. Background
3.2. Study selection and inclusion criteria
Lung cancer (LC) is the most common cause of cancer morbidity and mortality worldwide and in China. The pathogenesis of LC is believed to result from the interaction between environmental and genetic factors. Although smoking is the major etiologic factor, not all smokers develop LC, and approximately 25% of those with LC are never smokers [1]. Other etiologic factors, including passive smoking, occupational exposure to asbestos and certain metals, and air population have been implicated as contributing to the development of LC [2,3]. Candidate gene studies and genomewide association studies have revealed many dysfunctional genes that might be associated with LC [4]. However, these identified susceptibility factors do not fully explain the observed incidence of LC, indicating that there exist additional factors that also directly influence LC risk. HPV is believed to be an important factor in the pathogenesis of certain benign and malignant lesions in humans. Because its E6 and E7 oncoproteins are particularly proficient in disrupting normal epithelial differentiation and DNA synthesis by inhibiting the effective regulation of the cell cycle through inactive P53 and Rb proteins, high-risk HPV (HR-HPV) leads to genomic instability that can result in malignant transformation [5–8]. HR-HPV infection has been thought to be the main cause of human cervical cancers [9], a substantial proportion of other anogenital cancers [10,11] and oropharyngeal cancers [12–14]. While, HR-HPV as a contributor to LC was hypothesized many decades ago; a firm link between HPV and LC remains unclear. The first evidence that HPV might be involved in LC was provided by Syrjänen in 1979 and was based on a histological comparison of bronchial squamous cell carcinoma and the HPV infected female genital tract [15]. Syrjänen also conducted the first systematic review of HPV and LC and proposed that HPV might contribute to LC development [16]. Many studies have reported HPV prevalence rates in LCs that vary from 0.0% to 78.3% across divergent geographical regions and histological types of LC [17]. The largest systematic review by Syrjänen included 100 eligible studies with 7381 LCs that indicated HPV prevalence had reached 22.0% [3]. Although there have been numerous published studies regarding HPV prevalence, no systematic review or meta-analysis has quantified the association between LC patients with HPV infection and non-cancer controls. Understanding the association between HPV infection and LC could have important implications for cancer prevention, especially with respect to HPV vaccination policies for males.
The studies selected had to meet the following criteria: (1) case–control, cross-sectional or cohort studies compared HPV infection in lung tissue among LC patients and non-cancer controls; (2) HPV types were specified; (3) histological diagnosis of cases and controls were established; (4) HPV detection was based on DNA; (5) there were no restrictions based on patients’ nationality, ethnicity or gender; (6) there were no pediatric subjects included; (7) sufficient information was provided to calculate odds ratio (OR) with 95% confidence intervals (CI); (8) when an overlap of patients was found in several studies, only the study with the largest sample size and detailed information or the study that met the above criteria was included. Information was independently extracted from all eligible publications by two investigators, and discrepancies were resolved through discussion or via a third researcher.
2. Objectives The aim of this meta-analysis is to quantify the association between HPV infection in lung tissue and LC.
3.3. Data extraction An initial screening of the title and abstract was performed in the first step, followed by a further screening based on a fulltext review. All data were extracted independently and crosscheck by two investigators; disagreements were discussed until consensus or a third researcher was involved. For studies meeting our inclusion criteria, the following data were collected: first author, publication year, country of study, specimen type, histological type, HPV detection method, HPV types, and numbers of HPV positive and negative subjects in LCs and control groups. 3.4. Statistical analysis When sufficient data were available, a meta-analysis was performed and ORs with corresponding 95% CIs were calculated. Heterogeneity among studies was examined using the Cochran’s Q test by calculating the P value and quantified using the I2 statistic [19]. In this meta-analysis, if I2 < 50%, a fixed-effect model (Mantel–Haenszel method) was used to evaluate inter-study heterogeneity. Otherwise, a random-effect model (DerSimonian and Laird method) was used. Z test was used to determine the pooled OR and 95% CI. If necessary, a meta-regression analysis was conducted to identify the possible sources of heterogeneity. Subgroup analyses were conducted for ethnicity and different LC histological subtypes (SCC and AC). To evaluate the effect of one single study on the overall risk of LC, sensitivity analyses were performed by sequential omission of individual studies and the recalculating of pooled ORs and 95% CIs were performed. The potential influence of publication bias was assessed by visual inspection using Begg’s funnel plot and assessed statistically using Egger’s linear regression test [20,21]. All statistical tests were performed with the Stata 12.0 (StataCorporation, College Station, TX, USA). A 2-tailed P < 0.05 was considered statistically significant.
3. Study design
4. Results
3.1. Search strategy
4.1. Eligible studies
MEDLINE (PubMed), EMBASE (OVID) and Web of Science were searched to find relevant publications up to March 2014, using the search terms “lung cancer”, “human papillomavirus”, “HPV” and combinations thereof. The search was limited to studies that had been conducted on human subjects and written in English. Meeting abstracts were excluded because of limited data they offered. Reference lists of the retrieved articles, reviews and editorials were also screened to find all additional eligible studies. This meta-analysis was performed in accordance with PRISMA guidelines [18].
Systematic searching was performed to identify all relevant and eligible studies (Supplementary Fig. 1). Using the search terms, 287 potential studies were identified from three databases, and 195 studies were required for further assessment. After reviewing the full text of the studies, 184 were excluded because of inappropriate study design or control samples. When two studies were published by the same authors with the same patients [22,23], we selected the more recent and highest quality study [23]. Two papers used the same patients to genotype HPV6 and 11 in one study and
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
3
Table 1 Characteristics of the nine eligible studies in this meta-analysis. Study
Country
Sample
Li et al. (1995)
China
PET or FF
Cheng et al. (2001)
Taiwan
PET
Fei et al. (2006) Nadji et al. (2007)
China Iran
PET PET
Wang et al. (2008)
China
FF
Wang et al. (2010)
China
FF
Carpagnano et al. (2011)
Italy
FF and PET
Yu et al. (2013)a
China
PET
Sarchianaki et al. (2013)b
Greece
PET
Detection method
E6/E7 region PCR and DB L1 region MY09/MY11 and type-specific PCR ISH L1 region MY09/MY11 and GP5+/GP6+ PCR, sequencing E6/E7 region type-specific PCR E6/E7 region type-specific PCR HPV signc , pyrosequencing, INFINITI HPV-QUAD assayd SPF10 PCR and INNO-LIPA genotyping system and SB L1 region GP5+/GP6+ PCR and LINEAR ARRAY HPV Genotyping Test
HPV types detectable
HPV types detected
Cases (n = 1094)
Controls (n = 484)
HPV (+)
HPV (−)
HPV (+)
HPV (−)
16, 18
16, 18
16
34
0
18
16, 18
16, 18
77
64
16
44
16, 18 Unspecified
16, 18 6, 11, 26, 31, 16, 18
23 33
50 96
2 8
32 81
16, 18
16, 18
138
175
4
92
16, 18
16, 18
18
27
0
16
16, 18, 31, 33, 45, 35/68, 39/56, 58/52, 59/51, 6/11 16, 18
16, 30, 31, 39
12
61
0
68
16, 18
75
95
21
70
6, 11, 16, 18, 31, 33, 59
19
81
0
12
37 high- and low-risk types
Abbreviation: PET, paraffin-embedded tissue; FF, fresh frozen; PCR, polymerase chain reaction; DB, dot blot hybridization; ISH, in situ hybridization; SB, Southern blot hybridization. a INNO-LIPA genotyping system (Inno-genetics, Belgium) could assess 25 HPV types. In this study, SB was used to verify the PCR results in controversial cases. To avoid mistake, we used data that were consistent in these two methods. b In this study, real-time PCR using general primers (GP5+/GP6+) was performed to identify HPV positive cases. Second, HPV-positive samples were genotyped using LINEAR ARRAY HPV Genotyping Test (Roche). c Commercial assay for HPV genotyping (Diatech, Jesi, Italy). d Commercial assay to confirm HPV results (AutoGenomics, Carlsbad, CA, USA).
HPV16 and 18 in another study [24,25]. Because HPV16 and 18 are the main oncogenic HPVs, the study that genotyped HPV6 and 11 was excluded [25]. Following the strict inclusion criteria, nine studies were analyzed for HPV association in this meta-analysis. These studies were published between 1995 and 2013 and are listed in Table 1. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcv.2014.09.014.
to 409 (cases/controls: 313/96) [29]. In summary, 1578 subjects (cases/controls: 1094/484), 908 subjects (cases/controls: 611/297) and 767 subjects (cases/controls: 538/229) were enrolled to examine the association between infection with HPV16 and/or HPV18, HPV16 alone or HPV18 alone and LC risk (Table 2). Moreover, 529 SCCs [23,26–30], 224 ACs [23,26–29] and 255 non-cancer controls were analyzed to investigate the potential association between HPV16 and/or 18 and histological subtypes (Table 3).
4.2. Study characteristics
4.3. HPV infection and LC
The characteristics of the nine eligible studies, including country of origin, specimen type, HPV detection method, HPV type detected, histological type, and the number of LC subjects and non-cancer controls with and without HPV are presented in Tables 1 and 2. Among these, seven were conducted in Asia [23,24,26–30], and the remaining two were conducted in Europe [31,32]. Specimens included paraffin-embedded tissue and frozen biopsies. Furthermore, eight studies used a polymerase chain reaction (PCR)-based technique to detect HPV [23,24,26,28–32], while only one used in situ hybridization (ISH) [27]. Histological subtypes of LC tissues included SCC, AC, small cell lung cancer (SCLC), undifferentiated carcinoma and bronchioalveolar carcinoma. Non-cancer control tissues were diagnosed with benign pulmonary disease. Of the nine eligible studies, eight [23,24,27–32] presented control tissues in detail (pneumothorax, tuberculosis, chest wall deformity, cryptococcal infection, fibrosis, pseudotumor, bullae lung, lung cyst, chronic obstructive pulmonary disease with bronchial epithelial hyperplasia and squamous metaplasia, sarcoidosis, bronchial pneumonia, hamartomas, mycetomas, pulmonary abscess, endothoracic neurinoma, Klemperer’s tumor and typical carcinoid). The one exception [26] described the control tissues as “lung tissues of benign disease” with no further information. The sample sizes of the nine eligible studies ranged from 61 (cases/controls: 45/16) [30]
HPV (including low- and high-risk HPV) prevalence was higher among LC patients than among non-cancer controls (37.57% vs. 10.54%). Fig. 1 shows a forest plot of the overall association between subjects with/without HPV, stratified by ethnicity. A statistically significant association was observed between HPV and LC patients (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001; I2 = 53.3%). In addition, significance was also detected individually for the five Chinese studies (OR = 6.23, 95% CI: 2.78–13.97, P < 0.001, I2 = 66.7%) and four other studies (OR = 4.48, 95% CI: 1.82–11.01, P = 0.001, I2 = 6.5%). HPV incidence in Chinese (China and Taiwan) LC patients reached 43.81%. In LC and non-cancer controls, HPV16 prevalence was 19.80% and 3.37%, respectively. The pooled OR across studies was 5.84 (95% CI: 3.14–10.86, P < 0.001; I2 = 0.0%). Similar results were also observed in the HPV18 analysis. The HPV18 positive rate was 18.59% in LCs and 5.24% in non-cancer controls. The pooled estimate for the association with HPV18 was significant (OR = 4.29, 95% CI: 2.34–7.86, P < 0.001; I2 = 0.0%). The mean HPV16/18 incidence was 35.47% in LCs and 9.50% in non-cancer controls. The summary OR was 6.02 (95% CI: 3.22–11.28, P < 0.001; I2 = 50.7%) We also evaluated the cancer risk with HPV16/18 in different LC histological types. In SCC, HPV16/18 was significantly associated with cancer risk (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001, I2 = 44.9%); however, OR was not significant in AC (OR = 3.69, 95%
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
12 217
To evaluate the influence of each study on the pooled OR, we sequentially removed individual studies from the meta-analysis. The pooled ORs were stable and demonstrated statistical significance using the fixed-effect model, before and after deletion of any single study (data not shown). Together, these data indicate that the results of this meta-analysis are reliable and were not overly influenced by any one of the nine studies. Although the funnel plot shows some asymmetry (Fig. 2), Begg’s test suggested no significant publication bias existed in this meta-analysis (P = 0.076). However, the result of Egger’s test was significant (P = 0.047). Furthermore, the limited number of studies (n = 9) indicated a potential publication bias.
98 438 2 100
0 12
16 39 6
0
18 53 32 86
The number of individuals with HPV16/18 (+) included subjects with single infection (HPV16 or 18) and subjects multiple infections (HPV16 and 18). a
12 287 92 490 8 121
0 10
16 68 0 0 32 66 13 7
HPV18 (−) HPV18 (+)
0 7 2 3 44 83 56 118
HPV18 (−) HPV18(+)
6 58 17 11 18 51 33 89
HPV16 (−) HPV16 (+)
0 9 1 0 39 91 54 116
HPV16 (−)
11 50 19 13
HPV16(+) HPV16/18 (−)
18 44 32 86 92 16 68 70 12 438 0 16 2 3 4 0 0 21 0 46
HPV16/18 (+)
34 64 50 105 175 27 66 95 90 706
HPV16/18 (−) HPV16/18 (+)
4.4. Sensitivity analysis and publication bias
5. Discussion
16 77 23 24 138 18 7 75 10 388
Cases (n = 538) Controls (n = 297) Cases (n = 611) Controls (n = 484) Cases (n = 1094) Study
Table 2 Summary of nine eligible studies comparing lung cancer patients and non-cancer controls with/without HPV16/18 infection.a
CI: 0.99–13.71, P = 0.052; I2 = 75.5%). HPV16/18 prevalence in SCC and AC was 45.18% and 22.78%, respectively.
Li et al. (1995) Cheng et al. (2001) Fei et al. (2006) Nadji et al. (2007) Wang et al. (2008) Wang et al. (2010) Carpagnano et al. (2011) Yu et al. (2013) Sarchianaki et al. (2013) Total
Controls (n = 229)
4
Since Syrjänen first suggestion in 1979 that HPV may be associated with LC, several studies have explored this relationship. However, the results of these studies have often been controversial. HPV DNA has been identified in the serum, plasma and peripheral blood mononuclear cells of cervical cancers, head and neck cancers, LCs and healthy subjects [33–36]. These data indicate that blood circulation plays a vital role in HPV transmission. Based on many etiological studies, it is commonly agreed that HPV does not cause a generalized viremia. Therefore, assessment of the association between HPV and LC using different types of specimens may be inaccurate, particularly when using blood samples. Accordingly, studies using peripheral blood cells were excluded in this meta-analysis [36,37]. Most people are infected with HPV at some point in their lives, but only persistent infections will cause pathological changes. Studies detecting HPV by serology alone were also excluded, as this measure only indicates HPV exposure, not the exact site of infection [38]. LC tissues are more likely to be positive for HPV than normal lung tissues, which include tissues from LC adjacent or non-adjacent normal and from non-cancer controls (OR = 3.86, 95% CI: 2.87–5.19) [39]. However, we only included studies detecting HPV in lung tissue among LC patients and non-cancer controls that provided the most convincing evidence for association. This meta-analysis indicated a close association between HPV infection and LC. Thus, with strict case-control setting, these results suggest an increased risk and a strong association with HPV infected LCs compared with non-cancer controls (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001; I2 = 53.3%). Our study also quantifies that HPV16/18 is associated with LC (OR = 6.02, 95% CI: 3.32–11.28, P < 0.001) and significantly associated with SCC (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001). We propose that HPV plays a distinct role in the pathogenesis of different LCs. HPVs first infect the basal layer cells in the skin or mucosa. The infected cells then divide and laterally expand. Viral particles are released at the cell surface and infect addition tissues [40]. Metaplasia could be induced in the squamous epithelium, which could then transform to different subtypes of LC. It is reported that SCLC originating from neuroendocrine cells carry a very low prevalence of HPV [3]. In our analysis, the sample size of SCLC was too small [26,28,31] to determine the association between HPV and SCLC. The lack of statistical significance in the association between HPV and AC may also be due to the small number of AC cases (n = 224) included in the study. The lower prevalence of HPV association with adjacent normal lung tissues in HPV-positive LC [41,42] indicated that HPV likely plays an important role in lung carcinogenesis.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
5
Fig. 1. Overall association between HPV infection and lung cancer risk, stratified by ethnicity. For each study, the estimate of odds ratio (OR) and its 95% confidence intervals (CI) are plotted with a box and a horizontal line. The symbol diamond indicates pooled OR and its 95% CI. Eight studies used a PCR based method to assess HPV type [23,24,26,28–32], and one used ISH [27].
The relationship between HPV infection in LC and survival is still uncertain. Recently, Wang et al. (2014) described ACs with HPV16/18 infections as having significantly higher survival rates than those that are HPV16/18 negative [43]. Ragin et al. (2014) have reported that patients with HPV16/18 LCs had improved survival rates, although no significant differences were observed between the Kaplan–Meier curves. Several authors have reported that HPV viral load may be a useful predictor for developing cervical cancer precursors [44–46]. No data regarding HPV viral load and LC outcome or prognosis has yet been published. As measuring the viral load in living LC tissue is difficult, further studies are needed using larger sample sizes to elucidate the causal relationship of HPV infection and viral load to LC progression and to therapeutic response and survival. Limitations of our meta-analysis should be considered. First, several detection methods were used. While PCR-based methods and ISH have high specificity, ISH has low sensitivity and PCR has a higher rate of false positives [47]. Furthermore, long DNA fragments are known to be difficult to amplify from paraffin-embedded tissue. In the present analysis, six of the nine studies used this type of genomic DNA in HPV typing with different primers. Studies using consensus primers (MY09/MY11) to amplify the 450 bp fragment of the L1 gene may have lower sensitivity for HPV detection. Assessing primer factors for heterogeneity is difficult because many PCR primers were used in only nine studies. Given that a medium degree of heterogeneity [48] was detected in assessing the
Fig. 2. Begg’s funnel plot with pseudo 95% CI of publication bias on the association between HPV infection and LC risk.
association between LCs with/without HPV and non-cancer controls (I2 = 53.3%, P = 0.029), we performed a meta-regression analysis to explore the source of heterogeneity. Our results showed that ethnicity was the source of heterogeneity (P = 0.005).
Table 3 Summary of six eligible studies comparing lung squamous cell carcinoma and adenocarcinoma patients and non-cancer controls with/without HPV16/18 infection. Study
Li et al. (1995) Fei et al. (2006) Nadji et al. (2007) Wang et al. (2008) Wang et al. (2010) Yu et al. (2013) Total
SCCs (n = 529)
ACs (n = 224)
Controls (n = 344)
HPV16/18 (+)
HPV16/18 (−)
HPV16/18 (+)
13 13 19 112 18 64 239
14 27 76 103 27 43 290
2 10 2 26
HPV16/18 (−) 14 23 12 72
11 51
52 173
HPV16/18 (+)
HPV16/18 (−)
0 2 3 4 0 21 30
18 32 86 92 16 70 314
Abbreviation: SCCs, squamous cell carcinomas; ACs, adenocarcinomas.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
6
We contend that the small number of studies contributed to this type of between-study variability. Second, certain important characteristics have not been considered. A previous systematic review showed that smoking is strongly related in a dose-related manner with LC of all subtypes [49]. Studies also have shown higher percentages of HPV positivity in LC smokers than in ex- or non-smokers [25,28,29,31]. Whether smoking interacts with HPV to promote the development of LC is unclear. Our study was limited by the lack of available data regarding smoking status, gender and age. To more precisely determine the association between HPV and LC risk, these confounders must be adjusted as such analysis is not possible with the present data. The main limitation in the present study is the small number of included studies (n = 9). Generally, the sample size of controls is equal or more than that of cases. In this meta-analysis, the total number of cases is almost double that of controls. All of these could cause bias, although it could not be avoided. Many studies suggest that HPV prevalence in LC is related to ethnicity. A systematic review of 24 studies that included 911 patients showed that the HPV positive rate in Chinese LCs was 44.3% (fixed-effects model) or 37.7% (random-effects model) [3], which is close to our result (43.81%) and thus supports the accuracy of our analysis. China and Taiwan, other Asian countries and South America are HPV highincidence regions (HIR), while Australia, Europe, North America are HPV low-incidence regions (LIR) [3]. Because a small number of eligible case-control studies in LIRs were found, our study could not strictly stratify ethnicity for each region to quantify this association. A more detailed analysis and a larger number of representative subjects are needed. Furthermore, the association between HPV and ACs must also be verified. In summary, our meta-analysis indicates that HPV infection increases LC risk, particularly in SCC. Future studies should focus on the association between HPV and different histological LC subtypes using the same genotyping method and adjusting for more detailed demographic parameters. The pathogenesis and prognostic role of HPV in LC should be fully elucidated. Our study also suggests that HPV vaccination might be useful not only for cervical cancer prevention but also in the reduction of LC risk. Authors’ contribution Conceived and designed the experiments: K. Zhai and H.Z. Shi. Performed the experiments: K. Zhai and J. Ding. Analyzed the data: K. Zhai, J. Ding and H.Z. Shi. Wrote the paper: K. Zhai, J. Ding and H.Z. Shi. Funding This work was supported by National Natural Science Foundation of China (No. 81302052) and Natural Science Foundation of Beijing (No. 7142061). Competing interests None. Ethical approval Not required. References [1] Sun S, Schiller JH, Gazdar AF. Lung cancer in never smokers – a different disease. Nat Rev Cancer 2007;7:778–90. [2] Spyratos D, Zarogoulidis P, Porpodis K, Tsakiridis K, Machairiotis N, Katsikogiannis N, et al. Occupational exposure and lung cancer. J Thorac Dis 2013;5:S440–5.
[3] Syrjänen K. Detection of human papillomavirus in lung cancer: systematic review and meta-analysis. Anticancer Res 2012;32:3235–50. [4] Brennan P, Hainaut P, Boffetta P. Genetics of lung-cancer susceptibility. Lancet Oncol 2011;12:399–408. [5] Doorbar J. Papillomavirus life cycle organization and biomarker selection. Dis Markers 2007;23:297–313. [6] Munger K, Werness BA, Dyson N, Phelps WC, Harlow E, Howley PM. Complex formation of human papillomavirus E7 proteins with the retinoblastoma tumor suppressor gene product. EMBO J 1989;8:4099–105. [7] Werness BA, Levine AJ, Howley PM. Association of human papillomavirus types 16 and 18 E6 proteins with p53. Science 1990;248:76–9. [8] Stanley MA, Pett MR, Coleman N. HPV: from infection to cancer. Biochem Soc Trans 2007;35:1456–60. [9] Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9. [10] Steenbergen RD, de Wilde J, Wilting SM, Brink AA, Snijders PJ, Meijer CJ. HPVmediated transformation of the anogenital tract. J Clin Virol 2005;32(Suppl. 1):S25–33. [11] Heideman DA, Waterboer T, Pawlita M, Delis-van Diemen P, Nindl I, Leijte JA, et al. Human papillomavirus-16 is the predominant type etiologically involved in penile squamous cell carcinoma. J Clin Oncol 2007;25:4550–6. [12] Kreimer AR, Clifford GM, Boyle P, Franceschi S. Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review. Cancer Epidemiol Biomarkers Prev 2005;14:467–75. [13] Hobbs CG, Sterne JA, Bailey M, Heyderman RS, Birchall MA, Thomas SJ. Human papillomavirus and head and neck cancer: a systematic review and metaanalysis. Clin Otolaryngol 2006;31:259–66. [14] Syrjänen S, Lodi G, von Bultzingslowen I, Aliko A, Arduino P, Campisi G, et al. Human papillomaviruses in oral carcinoma and oral potentially malignant disorders: a systematic review. Oral Dis 2011;17(Suppl. 1):58–72. [15] Syrjänen KJ. Condylomatous changes in neoplastic bronchial epithelium. Report of a case. Respiration 1979;38:299–304. [16] Syrjänen KJ. HPV infections and lung cancer. J Clin Pathol 2002;55:885–91. [17] Srinivasan M, Taioli E, Ragin CC. Human papillomavirus type 16 and 18 in primary lung cancers – a meta-analysis. Carcinogenesis 2009;30: 1722–8. [18] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. [19] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58. [20] Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. [21] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34. [22] Yu Y, Yang A, Hu S, Yan H. Correlation of HPV-16/18 infection of human papillomavirus with lung squamous cell carcinomas in Western China. Oncol Rep 2009;21:1627–32. [23] Yu Y, Yang A, Hu S, Zhang J, Yan H. Significance of human papillomavirus 16/18 infection in association with p53 mutation in lung carcinomas. Clin Respir J 2013;7:27–33. [24] Cheng YW, Chiou HL, Sheu GT, Hsieh LL, Chen JT, Chen CY, et al. The association of human papillomavirus 16/18 infection with lung cancer among nonsmoking Taiwanese women. Cancer Res 2001;61:2799–803. [25] Cheng YW, Chiou HL, Chen JT, Chou MC, Lin TS, Lai WW, et al. Gender difference in human papillomarvirus infection for non-small cell lung cancer in Taiwan. Lung Cancer 2004;46:165–70. [26] Li Q, Hu K, Pan X, Cao Z, Yang J, Hu S. Detection of human papillomavirus types 16, 18 DNA related sequences in bronchogenic carcinoma by polymerase chain reaction. Chin Med J (Engl) 1995;108:610–4. [27] Fei Y, Yang J, Hsieh WC, Wu JY, Wu TC, Goan YG, et al. Different human papillomavirus 16/18 infection in Chinese non-small cell lung cancer patients living in Wuhan, China. Jpn J Clin Oncol 2006;36:274–9. [28] Nadji SA, Mokhtari-Azad T, Mahmoodi M, Yahyapour Y, Naghshvar F, Torabizadeh J, et al. Relationship between lung cancer and human papillomavirus in north of Iran, Mazandaran province. Cancer Lett 2007;248:41–6. [29] Wang Y, Wang A, Jiang R, Pan H, Huang B, Lu Y, et al. Human papillomavirus type 16 and 18 infection is associated with lung cancer patients from the central part of China. Oncol Rep 2008;20:333–9. [30] Wang YH, Chen DJ, Yi TN, Liu XH. The relationship among human papilloma virus infection, survivin, and p53 gene in lung squamous carcinoma tissue. Saudi Med J 2010;31:1331–6. [31] Carpagnano GE, Koutelou A, Natalicchio MI, Martinelli D, Ruggieri C, Di Taranto A, et al. HPV in exhaled breath condensate of lung cancer patients. Br J Cancer 2011;105:1183–90. [32] Sarchianaki E, Derdas SP, Ntaoukakis M, Vakonaki E, Lagoudaki ED, Lasithiotaki I, et al. Detection and genotype analysis of human papillomavirus in non-small cell lung cancer patients. Tumour Biol 2013, http://dx.doi.org/10.1007/s13277-013-1419-2. [33] Pao CC, Lin SS, Lin CY, Maa JS, Lai CH, Hsieh TT. Identification of human papillomavirus DNA sequences in peripheral blood mononuclear cells. Am J Clin Pathol 1991;95:540–6. [34] Liu VW, Tsang P, Yip A, Ng TY, Wong LC, Ngan HY. Low incidence of HPV DNA in sera of pretreatment cervical cancer patients. Gynecol Oncol 2001;82: 269–72.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
[35] Dong SM, Pai SI, Rha SH, Hildesheim A, Kurman RJ, Schwartz PE, et al. Detection and quantitation of human papillomavirus DNA in the plasma of patients with cervical carcinoma. Cancer Epidemiol Biomarkers Prev 2002;11: 3–6. [36] Chiou HL, Wu MF, Liaw YC, Cheng YW, Wong RH, Chen CY, et al. The presence of human papillomavirus type 16/18 DNA in blood circulation may act as a risk marker of lung cancer in Taiwan. Cancer 2003;97:1558–63. [37] Jain N, Singh V, Hedau S, Kumar S, Daga MK, Dewan R, et al. Infection of human papillomavirus type 18 and p53 codon 72 polymorphism in lung cancer patients from India. Chest 2005;128:3999–4007. [38] Simen-Kapeu A, Surcel HM, Koskela P, Pukkala E, Lehtinen M. Lack of association between human papillomavirus type 16 and 18 infections and female lung cancer. Cancer Epidemiol Biomarkers Prev 2010;19:1879–81. [39] Ragin C, Obikoya-Malomo M, Kim S, Chen Z, Flores-Obando R, Gibbs D, et al. HPV-associated lung cancers: an international pooled analysis. Carcinogenesis 2014, http://dx.doi.org/10.1093/carcin/bgu038. [40] zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application. Nat Rev Cancer 2002;2:342–50. [41] Krikelis D, Tzimagiorgis G, Georgiou E, Destouni C, Agorastos T, Haitoglou C, et al. Frequent presence of incomplete HPV16 E7 ORFs in lung carcinomas: memories of viral infection. J Clin Virol 2010;49:169–74. [42] Sagerup CM, Nymoen DA, Halvorsen AR, Lund-Iversen M, Helland A, Brustugun OT. Human papilloma virus detection and typing in 334 lung cancer patients. Acta Oncol 2014, http://dx.doi.org/10.3109/0284186X.2013.879608.
7
[43] Wang JL, Fang CL, Wang M, Yu MC, Bai KJ, Lu PC, et al. Human papillomavirus infections as a marker to predict overall survival in lung adenocarcinoma. Int J Cancer 2014;134:65–71. [44] Wu Y, Chen Y, Li L, Yu G, Zhang Y, He Y. Associations of high-risk HPV types and viral load with cervical cancer in China. J Clin Virol 2006;35: 264–9. [45] Gravitt PE, Kovacic MB, Herrero R, Schiffman M, Bratti C, Hildesheim A, et al. High load for most high risk human papillomavirus genotypes is associated with prevalent cervical cancer precursors but only HPV16 load predicts the development of incident disease. Int J Cancer 2007;121: 2787–93. [46] Huang Y, Huang MN, Li N, Li XG, Wu LY. Association between human papillomavirus DNA load and development of cervical intraepithelial neoplasia and cervical cancer. Int J Gynecol Cancer 2008;18:755–60. [47] Kimple AJ, Torres AD, Yang RZ, Kimple RJ. HPV-associated head and neck cancer: molecular and nano-scale markers for prognosis and therapeutic stratification. Sensors 2012;12:5159–69. [48] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. [49] Lee PN, Forey BA, Coombs KJ. Systematic review with meta-analysis of the epidemiological evidence in the 1900 relating smoking to lung cancer. BMC Cancer 2012;12:385.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
Journal of Clinical Virology xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Journal of Clinical Virology journal homepage: www.elsevier.com/locate/jcv
Review
HPV and lung cancer risk: A meta-analysis Kan Zhai a,∗,1 , Jie Ding a,1 , Huan-Zhong Shi a,b,∗ a b
Medical Research Center, Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China Department of Respiratory and Critical Care Medicine, Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China
a r t i c l e
i n f o
Article history: Received 30 April 2014 Received in revised form 21 September 2014 Accepted 22 September 2014 Keywords: HPV Lung cancer Risk Meta-analysis
a b s t r a c t The potential causal association between human papillomavirus (HPV) and lung cancer (LC) remains controversial. We performed this meta-analysis to evaluate whether HPV infection in lung tissue is associated with LC compared with non-cancer controls. We also quantified this association in different LC subtypes. MEDLINE, EMBASE and Web of Science were searched through March 2014, using the search terms “lung cancer”, “human papillomavirus”, “HPV” and their combinations. Association was tested using odds ratio (OR) with 95% confidence intervals (95% CI). Heterogeneity was assessed using Q and I2 statistic. Finally, nine studies, for a total of 1094 LCs and 484 non-cancer controls, were identified as eligible publications. The pooled results showed that HPV infection was associated with LC (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001). Similar results were also observed in HPV16 and/or HPV18 (HPV16/18) infection analyses (OR = 6.02, 95% CI: 3.22–11.28, P < 0.001). HPV16/18 was significantly associated with lung squamous cell carcinoma (SCC) (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001), while the pooled OR was 3.69 in lung adenocarcinoma (95% CI: 0.99–13.71, P = 0.052). Our results suggest that lung tissue with HPV infection has a strong association with LC, and especially, HPV16/18 infection significantly increases SCC risk, which indicates a potential pathogenesis link between HPV and LC. © 2014 Elsevier B.V. All rights reserved.
Contents 1. 2. 3.
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Study design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Search strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. 3.2. Study selection and inclusion criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3. Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4. Statistical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Eligible studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Study characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. HPV infection and LC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Sensitivity analysis and publication bias . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Discussion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Authors’ contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Competing interests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ethical approval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
∗ Corresponding authors at: Beijing Chao-Yang Hospital, Capital Medical University, Beijing 100020, China. Tel.: +86 10 85231624. E-mail addresses: kan [email protected] (K. Zhai), [email protected] (J. Ding), [email protected] (H.-Z. Shi). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.jcv.2014.09.014 1386-6532/© 2014 Elsevier B.V. All rights reserved.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
2
1. Background
3.2. Study selection and inclusion criteria
Lung cancer (LC) is the most common cause of cancer morbidity and mortality worldwide and in China. The pathogenesis of LC is believed to result from the interaction between environmental and genetic factors. Although smoking is the major etiologic factor, not all smokers develop LC, and approximately 25% of those with LC are never smokers [1]. Other etiologic factors, including passive smoking, occupational exposure to asbestos and certain metals, and air population have been implicated as contributing to the development of LC [2,3]. Candidate gene studies and genomewide association studies have revealed many dysfunctional genes that might be associated with LC [4]. However, these identified susceptibility factors do not fully explain the observed incidence of LC, indicating that there exist additional factors that also directly influence LC risk. HPV is believed to be an important factor in the pathogenesis of certain benign and malignant lesions in humans. Because its E6 and E7 oncoproteins are particularly proficient in disrupting normal epithelial differentiation and DNA synthesis by inhibiting the effective regulation of the cell cycle through inactive P53 and Rb proteins, high-risk HPV (HR-HPV) leads to genomic instability that can result in malignant transformation [5–8]. HR-HPV infection has been thought to be the main cause of human cervical cancers [9], a substantial proportion of other anogenital cancers [10,11] and oropharyngeal cancers [12–14]. While, HR-HPV as a contributor to LC was hypothesized many decades ago; a firm link between HPV and LC remains unclear. The first evidence that HPV might be involved in LC was provided by Syrjänen in 1979 and was based on a histological comparison of bronchial squamous cell carcinoma and the HPV infected female genital tract [15]. Syrjänen also conducted the first systematic review of HPV and LC and proposed that HPV might contribute to LC development [16]. Many studies have reported HPV prevalence rates in LCs that vary from 0.0% to 78.3% across divergent geographical regions and histological types of LC [17]. The largest systematic review by Syrjänen included 100 eligible studies with 7381 LCs that indicated HPV prevalence had reached 22.0% [3]. Although there have been numerous published studies regarding HPV prevalence, no systematic review or meta-analysis has quantified the association between LC patients with HPV infection and non-cancer controls. Understanding the association between HPV infection and LC could have important implications for cancer prevention, especially with respect to HPV vaccination policies for males.
The studies selected had to meet the following criteria: (1) case–control, cross-sectional or cohort studies compared HPV infection in lung tissue among LC patients and non-cancer controls; (2) HPV types were specified; (3) histological diagnosis of cases and controls were established; (4) HPV detection was based on DNA; (5) there were no restrictions based on patients’ nationality, ethnicity or gender; (6) there were no pediatric subjects included; (7) sufficient information was provided to calculate odds ratio (OR) with 95% confidence intervals (CI); (8) when an overlap of patients was found in several studies, only the study with the largest sample size and detailed information or the study that met the above criteria was included. Information was independently extracted from all eligible publications by two investigators, and discrepancies were resolved through discussion or via a third researcher.
2. Objectives The aim of this meta-analysis is to quantify the association between HPV infection in lung tissue and LC.
3.3. Data extraction An initial screening of the title and abstract was performed in the first step, followed by a further screening based on a fulltext review. All data were extracted independently and crosscheck by two investigators; disagreements were discussed until consensus or a third researcher was involved. For studies meeting our inclusion criteria, the following data were collected: first author, publication year, country of study, specimen type, histological type, HPV detection method, HPV types, and numbers of HPV positive and negative subjects in LCs and control groups. 3.4. Statistical analysis When sufficient data were available, a meta-analysis was performed and ORs with corresponding 95% CIs were calculated. Heterogeneity among studies was examined using the Cochran’s Q test by calculating the P value and quantified using the I2 statistic [19]. In this meta-analysis, if I2 < 50%, a fixed-effect model (Mantel–Haenszel method) was used to evaluate inter-study heterogeneity. Otherwise, a random-effect model (DerSimonian and Laird method) was used. Z test was used to determine the pooled OR and 95% CI. If necessary, a meta-regression analysis was conducted to identify the possible sources of heterogeneity. Subgroup analyses were conducted for ethnicity and different LC histological subtypes (SCC and AC). To evaluate the effect of one single study on the overall risk of LC, sensitivity analyses were performed by sequential omission of individual studies and the recalculating of pooled ORs and 95% CIs were performed. The potential influence of publication bias was assessed by visual inspection using Begg’s funnel plot and assessed statistically using Egger’s linear regression test [20,21]. All statistical tests were performed with the Stata 12.0 (StataCorporation, College Station, TX, USA). A 2-tailed P < 0.05 was considered statistically significant.
3. Study design
4. Results
3.1. Search strategy
4.1. Eligible studies
MEDLINE (PubMed), EMBASE (OVID) and Web of Science were searched to find relevant publications up to March 2014, using the search terms “lung cancer”, “human papillomavirus”, “HPV” and combinations thereof. The search was limited to studies that had been conducted on human subjects and written in English. Meeting abstracts were excluded because of limited data they offered. Reference lists of the retrieved articles, reviews and editorials were also screened to find all additional eligible studies. This meta-analysis was performed in accordance with PRISMA guidelines [18].
Systematic searching was performed to identify all relevant and eligible studies (Supplementary Fig. 1). Using the search terms, 287 potential studies were identified from three databases, and 195 studies were required for further assessment. After reviewing the full text of the studies, 184 were excluded because of inappropriate study design or control samples. When two studies were published by the same authors with the same patients [22,23], we selected the more recent and highest quality study [23]. Two papers used the same patients to genotype HPV6 and 11 in one study and
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
3
Table 1 Characteristics of the nine eligible studies in this meta-analysis. Study
Country
Sample
Li et al. (1995)
China
PET or FF
Cheng et al. (2001)
Taiwan
PET
Fei et al. (2006) Nadji et al. (2007)
China Iran
PET PET
Wang et al. (2008)
China
FF
Wang et al. (2010)
China
FF
Carpagnano et al. (2011)
Italy
FF and PET
Yu et al. (2013)a
China
PET
Sarchianaki et al. (2013)b
Greece
PET
Detection method
E6/E7 region PCR and DB L1 region MY09/MY11 and type-specific PCR ISH L1 region MY09/MY11 and GP5+/GP6+ PCR, sequencing E6/E7 region type-specific PCR E6/E7 region type-specific PCR HPV signc , pyrosequencing, INFINITI HPV-QUAD assayd SPF10 PCR and INNO-LIPA genotyping system and SB L1 region GP5+/GP6+ PCR and LINEAR ARRAY HPV Genotyping Test
HPV types detectable
HPV types detected
Cases (n = 1094)
Controls (n = 484)
HPV (+)
HPV (−)
HPV (+)
HPV (−)
16, 18
16, 18
16
34
0
18
16, 18
16, 18
77
64
16
44
16, 18 Unspecified
16, 18 6, 11, 26, 31, 16, 18
23 33
50 96
2 8
32 81
16, 18
16, 18
138
175
4
92
16, 18
16, 18
18
27
0
16
16, 18, 31, 33, 45, 35/68, 39/56, 58/52, 59/51, 6/11 16, 18
16, 30, 31, 39
12
61
0
68
16, 18
75
95
21
70
6, 11, 16, 18, 31, 33, 59
19
81
0
12
37 high- and low-risk types
Abbreviation: PET, paraffin-embedded tissue; FF, fresh frozen; PCR, polymerase chain reaction; DB, dot blot hybridization; ISH, in situ hybridization; SB, Southern blot hybridization. a INNO-LIPA genotyping system (Inno-genetics, Belgium) could assess 25 HPV types. In this study, SB was used to verify the PCR results in controversial cases. To avoid mistake, we used data that were consistent in these two methods. b In this study, real-time PCR using general primers (GP5+/GP6+) was performed to identify HPV positive cases. Second, HPV-positive samples were genotyped using LINEAR ARRAY HPV Genotyping Test (Roche). c Commercial assay for HPV genotyping (Diatech, Jesi, Italy). d Commercial assay to confirm HPV results (AutoGenomics, Carlsbad, CA, USA).
HPV16 and 18 in another study [24,25]. Because HPV16 and 18 are the main oncogenic HPVs, the study that genotyped HPV6 and 11 was excluded [25]. Following the strict inclusion criteria, nine studies were analyzed for HPV association in this meta-analysis. These studies were published between 1995 and 2013 and are listed in Table 1. Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcv.2014.09.014.
to 409 (cases/controls: 313/96) [29]. In summary, 1578 subjects (cases/controls: 1094/484), 908 subjects (cases/controls: 611/297) and 767 subjects (cases/controls: 538/229) were enrolled to examine the association between infection with HPV16 and/or HPV18, HPV16 alone or HPV18 alone and LC risk (Table 2). Moreover, 529 SCCs [23,26–30], 224 ACs [23,26–29] and 255 non-cancer controls were analyzed to investigate the potential association between HPV16 and/or 18 and histological subtypes (Table 3).
4.2. Study characteristics
4.3. HPV infection and LC
The characteristics of the nine eligible studies, including country of origin, specimen type, HPV detection method, HPV type detected, histological type, and the number of LC subjects and non-cancer controls with and without HPV are presented in Tables 1 and 2. Among these, seven were conducted in Asia [23,24,26–30], and the remaining two were conducted in Europe [31,32]. Specimens included paraffin-embedded tissue and frozen biopsies. Furthermore, eight studies used a polymerase chain reaction (PCR)-based technique to detect HPV [23,24,26,28–32], while only one used in situ hybridization (ISH) [27]. Histological subtypes of LC tissues included SCC, AC, small cell lung cancer (SCLC), undifferentiated carcinoma and bronchioalveolar carcinoma. Non-cancer control tissues were diagnosed with benign pulmonary disease. Of the nine eligible studies, eight [23,24,27–32] presented control tissues in detail (pneumothorax, tuberculosis, chest wall deformity, cryptococcal infection, fibrosis, pseudotumor, bullae lung, lung cyst, chronic obstructive pulmonary disease with bronchial epithelial hyperplasia and squamous metaplasia, sarcoidosis, bronchial pneumonia, hamartomas, mycetomas, pulmonary abscess, endothoracic neurinoma, Klemperer’s tumor and typical carcinoid). The one exception [26] described the control tissues as “lung tissues of benign disease” with no further information. The sample sizes of the nine eligible studies ranged from 61 (cases/controls: 45/16) [30]
HPV (including low- and high-risk HPV) prevalence was higher among LC patients than among non-cancer controls (37.57% vs. 10.54%). Fig. 1 shows a forest plot of the overall association between subjects with/without HPV, stratified by ethnicity. A statistically significant association was observed between HPV and LC patients (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001; I2 = 53.3%). In addition, significance was also detected individually for the five Chinese studies (OR = 6.23, 95% CI: 2.78–13.97, P < 0.001, I2 = 66.7%) and four other studies (OR = 4.48, 95% CI: 1.82–11.01, P = 0.001, I2 = 6.5%). HPV incidence in Chinese (China and Taiwan) LC patients reached 43.81%. In LC and non-cancer controls, HPV16 prevalence was 19.80% and 3.37%, respectively. The pooled OR across studies was 5.84 (95% CI: 3.14–10.86, P < 0.001; I2 = 0.0%). Similar results were also observed in the HPV18 analysis. The HPV18 positive rate was 18.59% in LCs and 5.24% in non-cancer controls. The pooled estimate for the association with HPV18 was significant (OR = 4.29, 95% CI: 2.34–7.86, P < 0.001; I2 = 0.0%). The mean HPV16/18 incidence was 35.47% in LCs and 9.50% in non-cancer controls. The summary OR was 6.02 (95% CI: 3.22–11.28, P < 0.001; I2 = 50.7%) We also evaluated the cancer risk with HPV16/18 in different LC histological types. In SCC, HPV16/18 was significantly associated with cancer risk (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001, I2 = 44.9%); however, OR was not significant in AC (OR = 3.69, 95%
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
12 217
To evaluate the influence of each study on the pooled OR, we sequentially removed individual studies from the meta-analysis. The pooled ORs were stable and demonstrated statistical significance using the fixed-effect model, before and after deletion of any single study (data not shown). Together, these data indicate that the results of this meta-analysis are reliable and were not overly influenced by any one of the nine studies. Although the funnel plot shows some asymmetry (Fig. 2), Begg’s test suggested no significant publication bias existed in this meta-analysis (P = 0.076). However, the result of Egger’s test was significant (P = 0.047). Furthermore, the limited number of studies (n = 9) indicated a potential publication bias.
98 438 2 100
0 12
16 39 6
0
18 53 32 86
The number of individuals with HPV16/18 (+) included subjects with single infection (HPV16 or 18) and subjects multiple infections (HPV16 and 18). a
12 287 92 490 8 121
0 10
16 68 0 0 32 66 13 7
HPV18 (−) HPV18 (+)
0 7 2 3 44 83 56 118
HPV18 (−) HPV18(+)
6 58 17 11 18 51 33 89
HPV16 (−) HPV16 (+)
0 9 1 0 39 91 54 116
HPV16 (−)
11 50 19 13
HPV16(+) HPV16/18 (−)
18 44 32 86 92 16 68 70 12 438 0 16 2 3 4 0 0 21 0 46
HPV16/18 (+)
34 64 50 105 175 27 66 95 90 706
HPV16/18 (−) HPV16/18 (+)
4.4. Sensitivity analysis and publication bias
5. Discussion
16 77 23 24 138 18 7 75 10 388
Cases (n = 538) Controls (n = 297) Cases (n = 611) Controls (n = 484) Cases (n = 1094) Study
Table 2 Summary of nine eligible studies comparing lung cancer patients and non-cancer controls with/without HPV16/18 infection.a
CI: 0.99–13.71, P = 0.052; I2 = 75.5%). HPV16/18 prevalence in SCC and AC was 45.18% and 22.78%, respectively.
Li et al. (1995) Cheng et al. (2001) Fei et al. (2006) Nadji et al. (2007) Wang et al. (2008) Wang et al. (2010) Carpagnano et al. (2011) Yu et al. (2013) Sarchianaki et al. (2013) Total
Controls (n = 229)
4
Since Syrjänen first suggestion in 1979 that HPV may be associated with LC, several studies have explored this relationship. However, the results of these studies have often been controversial. HPV DNA has been identified in the serum, plasma and peripheral blood mononuclear cells of cervical cancers, head and neck cancers, LCs and healthy subjects [33–36]. These data indicate that blood circulation plays a vital role in HPV transmission. Based on many etiological studies, it is commonly agreed that HPV does not cause a generalized viremia. Therefore, assessment of the association between HPV and LC using different types of specimens may be inaccurate, particularly when using blood samples. Accordingly, studies using peripheral blood cells were excluded in this meta-analysis [36,37]. Most people are infected with HPV at some point in their lives, but only persistent infections will cause pathological changes. Studies detecting HPV by serology alone were also excluded, as this measure only indicates HPV exposure, not the exact site of infection [38]. LC tissues are more likely to be positive for HPV than normal lung tissues, which include tissues from LC adjacent or non-adjacent normal and from non-cancer controls (OR = 3.86, 95% CI: 2.87–5.19) [39]. However, we only included studies detecting HPV in lung tissue among LC patients and non-cancer controls that provided the most convincing evidence for association. This meta-analysis indicated a close association between HPV infection and LC. Thus, with strict case-control setting, these results suggest an increased risk and a strong association with HPV infected LCs compared with non-cancer controls (OR = 5.67, 95% CI: 3.09–10.40, P < 0.001; I2 = 53.3%). Our study also quantifies that HPV16/18 is associated with LC (OR = 6.02, 95% CI: 3.32–11.28, P < 0.001) and significantly associated with SCC (OR = 9.78, 95% CI: 6.28–15.22, P < 0.001). We propose that HPV plays a distinct role in the pathogenesis of different LCs. HPVs first infect the basal layer cells in the skin or mucosa. The infected cells then divide and laterally expand. Viral particles are released at the cell surface and infect addition tissues [40]. Metaplasia could be induced in the squamous epithelium, which could then transform to different subtypes of LC. It is reported that SCLC originating from neuroendocrine cells carry a very low prevalence of HPV [3]. In our analysis, the sample size of SCLC was too small [26,28,31] to determine the association between HPV and SCLC. The lack of statistical significance in the association between HPV and AC may also be due to the small number of AC cases (n = 224) included in the study. The lower prevalence of HPV association with adjacent normal lung tissues in HPV-positive LC [41,42] indicated that HPV likely plays an important role in lung carcinogenesis.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model
ARTICLE IN PRESS
JCV-3137; No. of Pages 7
K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
5
Fig. 1. Overall association between HPV infection and lung cancer risk, stratified by ethnicity. For each study, the estimate of odds ratio (OR) and its 95% confidence intervals (CI) are plotted with a box and a horizontal line. The symbol diamond indicates pooled OR and its 95% CI. Eight studies used a PCR based method to assess HPV type [23,24,26,28–32], and one used ISH [27].
The relationship between HPV infection in LC and survival is still uncertain. Recently, Wang et al. (2014) described ACs with HPV16/18 infections as having significantly higher survival rates than those that are HPV16/18 negative [43]. Ragin et al. (2014) have reported that patients with HPV16/18 LCs had improved survival rates, although no significant differences were observed between the Kaplan–Meier curves. Several authors have reported that HPV viral load may be a useful predictor for developing cervical cancer precursors [44–46]. No data regarding HPV viral load and LC outcome or prognosis has yet been published. As measuring the viral load in living LC tissue is difficult, further studies are needed using larger sample sizes to elucidate the causal relationship of HPV infection and viral load to LC progression and to therapeutic response and survival. Limitations of our meta-analysis should be considered. First, several detection methods were used. While PCR-based methods and ISH have high specificity, ISH has low sensitivity and PCR has a higher rate of false positives [47]. Furthermore, long DNA fragments are known to be difficult to amplify from paraffin-embedded tissue. In the present analysis, six of the nine studies used this type of genomic DNA in HPV typing with different primers. Studies using consensus primers (MY09/MY11) to amplify the 450 bp fragment of the L1 gene may have lower sensitivity for HPV detection. Assessing primer factors for heterogeneity is difficult because many PCR primers were used in only nine studies. Given that a medium degree of heterogeneity [48] was detected in assessing the
Fig. 2. Begg’s funnel plot with pseudo 95% CI of publication bias on the association between HPV infection and LC risk.
association between LCs with/without HPV and non-cancer controls (I2 = 53.3%, P = 0.029), we performed a meta-regression analysis to explore the source of heterogeneity. Our results showed that ethnicity was the source of heterogeneity (P = 0.005).
Table 3 Summary of six eligible studies comparing lung squamous cell carcinoma and adenocarcinoma patients and non-cancer controls with/without HPV16/18 infection. Study
Li et al. (1995) Fei et al. (2006) Nadji et al. (2007) Wang et al. (2008) Wang et al. (2010) Yu et al. (2013) Total
SCCs (n = 529)
ACs (n = 224)
Controls (n = 344)
HPV16/18 (+)
HPV16/18 (−)
HPV16/18 (+)
13 13 19 112 18 64 239
14 27 76 103 27 43 290
2 10 2 26
HPV16/18 (−) 14 23 12 72
11 51
52 173
HPV16/18 (+)
HPV16/18 (−)
0 2 3 4 0 21 30
18 32 86 92 16 70 314
Abbreviation: SCCs, squamous cell carcinomas; ACs, adenocarcinomas.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
6
We contend that the small number of studies contributed to this type of between-study variability. Second, certain important characteristics have not been considered. A previous systematic review showed that smoking is strongly related in a dose-related manner with LC of all subtypes [49]. Studies also have shown higher percentages of HPV positivity in LC smokers than in ex- or non-smokers [25,28,29,31]. Whether smoking interacts with HPV to promote the development of LC is unclear. Our study was limited by the lack of available data regarding smoking status, gender and age. To more precisely determine the association between HPV and LC risk, these confounders must be adjusted as such analysis is not possible with the present data. The main limitation in the present study is the small number of included studies (n = 9). Generally, the sample size of controls is equal or more than that of cases. In this meta-analysis, the total number of cases is almost double that of controls. All of these could cause bias, although it could not be avoided. Many studies suggest that HPV prevalence in LC is related to ethnicity. A systematic review of 24 studies that included 911 patients showed that the HPV positive rate in Chinese LCs was 44.3% (fixed-effects model) or 37.7% (random-effects model) [3], which is close to our result (43.81%) and thus supports the accuracy of our analysis. China and Taiwan, other Asian countries and South America are HPV highincidence regions (HIR), while Australia, Europe, North America are HPV low-incidence regions (LIR) [3]. Because a small number of eligible case-control studies in LIRs were found, our study could not strictly stratify ethnicity for each region to quantify this association. A more detailed analysis and a larger number of representative subjects are needed. Furthermore, the association between HPV and ACs must also be verified. In summary, our meta-analysis indicates that HPV infection increases LC risk, particularly in SCC. Future studies should focus on the association between HPV and different histological LC subtypes using the same genotyping method and adjusting for more detailed demographic parameters. The pathogenesis and prognostic role of HPV in LC should be fully elucidated. Our study also suggests that HPV vaccination might be useful not only for cervical cancer prevention but also in the reduction of LC risk. Authors’ contribution Conceived and designed the experiments: K. Zhai and H.Z. Shi. Performed the experiments: K. Zhai and J. Ding. Analyzed the data: K. Zhai, J. Ding and H.Z. Shi. Wrote the paper: K. Zhai, J. Ding and H.Z. Shi. Funding This work was supported by National Natural Science Foundation of China (No. 81302052) and Natural Science Foundation of Beijing (No. 7142061). Competing interests None. Ethical approval Not required. References [1] Sun S, Schiller JH, Gazdar AF. Lung cancer in never smokers – a different disease. Nat Rev Cancer 2007;7:778–90. [2] Spyratos D, Zarogoulidis P, Porpodis K, Tsakiridis K, Machairiotis N, Katsikogiannis N, et al. Occupational exposure and lung cancer. J Thorac Dis 2013;5:S440–5.
[3] Syrjänen K. Detection of human papillomavirus in lung cancer: systematic review and meta-analysis. Anticancer Res 2012;32:3235–50. [4] Brennan P, Hainaut P, Boffetta P. Genetics of lung-cancer susceptibility. Lancet Oncol 2011;12:399–408. [5] Doorbar J. Papillomavirus life cycle organization and biomarker selection. Dis Markers 2007;23:297–313. [6] Munger K, Werness BA, Dyson N, Phelps WC, Harlow E, Howley PM. Complex formation of human papillomavirus E7 proteins with the retinoblastoma tumor suppressor gene product. EMBO J 1989;8:4099–105. [7] Werness BA, Levine AJ, Howley PM. Association of human papillomavirus types 16 and 18 E6 proteins with p53. Science 1990;248:76–9. [8] Stanley MA, Pett MR, Coleman N. HPV: from infection to cancer. Biochem Soc Trans 2007;35:1456–60. [9] Walboomers JM, Jacobs MV, Manos MM, Bosch FX, Kummer JA, Shah KV, et al. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 1999;189:12–9. [10] Steenbergen RD, de Wilde J, Wilting SM, Brink AA, Snijders PJ, Meijer CJ. HPVmediated transformation of the anogenital tract. J Clin Virol 2005;32(Suppl. 1):S25–33. [11] Heideman DA, Waterboer T, Pawlita M, Delis-van Diemen P, Nindl I, Leijte JA, et al. Human papillomavirus-16 is the predominant type etiologically involved in penile squamous cell carcinoma. J Clin Oncol 2007;25:4550–6. [12] Kreimer AR, Clifford GM, Boyle P, Franceschi S. Human papillomavirus types in head and neck squamous cell carcinomas worldwide: a systematic review. Cancer Epidemiol Biomarkers Prev 2005;14:467–75. [13] Hobbs CG, Sterne JA, Bailey M, Heyderman RS, Birchall MA, Thomas SJ. Human papillomavirus and head and neck cancer: a systematic review and metaanalysis. Clin Otolaryngol 2006;31:259–66. [14] Syrjänen S, Lodi G, von Bultzingslowen I, Aliko A, Arduino P, Campisi G, et al. Human papillomaviruses in oral carcinoma and oral potentially malignant disorders: a systematic review. Oral Dis 2011;17(Suppl. 1):58–72. [15] Syrjänen KJ. Condylomatous changes in neoplastic bronchial epithelium. Report of a case. Respiration 1979;38:299–304. [16] Syrjänen KJ. HPV infections and lung cancer. J Clin Pathol 2002;55:885–91. [17] Srinivasan M, Taioli E, Ragin CC. Human papillomavirus type 16 and 18 in primary lung cancers – a meta-analysis. Carcinogenesis 2009;30: 1722–8. [18] Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. [19] Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58. [20] Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101. [21] Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629–34. [22] Yu Y, Yang A, Hu S, Yan H. Correlation of HPV-16/18 infection of human papillomavirus with lung squamous cell carcinomas in Western China. Oncol Rep 2009;21:1627–32. [23] Yu Y, Yang A, Hu S, Zhang J, Yan H. Significance of human papillomavirus 16/18 infection in association with p53 mutation in lung carcinomas. Clin Respir J 2013;7:27–33. [24] Cheng YW, Chiou HL, Sheu GT, Hsieh LL, Chen JT, Chen CY, et al. The association of human papillomavirus 16/18 infection with lung cancer among nonsmoking Taiwanese women. Cancer Res 2001;61:2799–803. [25] Cheng YW, Chiou HL, Chen JT, Chou MC, Lin TS, Lai WW, et al. Gender difference in human papillomarvirus infection for non-small cell lung cancer in Taiwan. Lung Cancer 2004;46:165–70. [26] Li Q, Hu K, Pan X, Cao Z, Yang J, Hu S. Detection of human papillomavirus types 16, 18 DNA related sequences in bronchogenic carcinoma by polymerase chain reaction. Chin Med J (Engl) 1995;108:610–4. [27] Fei Y, Yang J, Hsieh WC, Wu JY, Wu TC, Goan YG, et al. Different human papillomavirus 16/18 infection in Chinese non-small cell lung cancer patients living in Wuhan, China. Jpn J Clin Oncol 2006;36:274–9. [28] Nadji SA, Mokhtari-Azad T, Mahmoodi M, Yahyapour Y, Naghshvar F, Torabizadeh J, et al. Relationship between lung cancer and human papillomavirus in north of Iran, Mazandaran province. Cancer Lett 2007;248:41–6. [29] Wang Y, Wang A, Jiang R, Pan H, Huang B, Lu Y, et al. Human papillomavirus type 16 and 18 infection is associated with lung cancer patients from the central part of China. Oncol Rep 2008;20:333–9. [30] Wang YH, Chen DJ, Yi TN, Liu XH. The relationship among human papilloma virus infection, survivin, and p53 gene in lung squamous carcinoma tissue. Saudi Med J 2010;31:1331–6. [31] Carpagnano GE, Koutelou A, Natalicchio MI, Martinelli D, Ruggieri C, Di Taranto A, et al. HPV in exhaled breath condensate of lung cancer patients. Br J Cancer 2011;105:1183–90. [32] Sarchianaki E, Derdas SP, Ntaoukakis M, Vakonaki E, Lagoudaki ED, Lasithiotaki I, et al. Detection and genotype analysis of human papillomavirus in non-small cell lung cancer patients. Tumour Biol 2013, http://dx.doi.org/10.1007/s13277-013-1419-2. [33] Pao CC, Lin SS, Lin CY, Maa JS, Lai CH, Hsieh TT. Identification of human papillomavirus DNA sequences in peripheral blood mononuclear cells. Am J Clin Pathol 1991;95:540–6. [34] Liu VW, Tsang P, Yip A, Ng TY, Wong LC, Ngan HY. Low incidence of HPV DNA in sera of pretreatment cervical cancer patients. Gynecol Oncol 2001;82: 269–72.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
G Model JCV-3137; No. of Pages 7
ARTICLE IN PRESS K. Zhai et al. / Journal of Clinical Virology xxx (2014) xxx–xxx
[35] Dong SM, Pai SI, Rha SH, Hildesheim A, Kurman RJ, Schwartz PE, et al. Detection and quantitation of human papillomavirus DNA in the plasma of patients with cervical carcinoma. Cancer Epidemiol Biomarkers Prev 2002;11: 3–6. [36] Chiou HL, Wu MF, Liaw YC, Cheng YW, Wong RH, Chen CY, et al. The presence of human papillomavirus type 16/18 DNA in blood circulation may act as a risk marker of lung cancer in Taiwan. Cancer 2003;97:1558–63. [37] Jain N, Singh V, Hedau S, Kumar S, Daga MK, Dewan R, et al. Infection of human papillomavirus type 18 and p53 codon 72 polymorphism in lung cancer patients from India. Chest 2005;128:3999–4007. [38] Simen-Kapeu A, Surcel HM, Koskela P, Pukkala E, Lehtinen M. Lack of association between human papillomavirus type 16 and 18 infections and female lung cancer. Cancer Epidemiol Biomarkers Prev 2010;19:1879–81. [39] Ragin C, Obikoya-Malomo M, Kim S, Chen Z, Flores-Obando R, Gibbs D, et al. HPV-associated lung cancers: an international pooled analysis. Carcinogenesis 2014, http://dx.doi.org/10.1093/carcin/bgu038. [40] zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application. Nat Rev Cancer 2002;2:342–50. [41] Krikelis D, Tzimagiorgis G, Georgiou E, Destouni C, Agorastos T, Haitoglou C, et al. Frequent presence of incomplete HPV16 E7 ORFs in lung carcinomas: memories of viral infection. J Clin Virol 2010;49:169–74. [42] Sagerup CM, Nymoen DA, Halvorsen AR, Lund-Iversen M, Helland A, Brustugun OT. Human papilloma virus detection and typing in 334 lung cancer patients. Acta Oncol 2014, http://dx.doi.org/10.3109/0284186X.2013.879608.
7
[43] Wang JL, Fang CL, Wang M, Yu MC, Bai KJ, Lu PC, et al. Human papillomavirus infections as a marker to predict overall survival in lung adenocarcinoma. Int J Cancer 2014;134:65–71. [44] Wu Y, Chen Y, Li L, Yu G, Zhang Y, He Y. Associations of high-risk HPV types and viral load with cervical cancer in China. J Clin Virol 2006;35: 264–9. [45] Gravitt PE, Kovacic MB, Herrero R, Schiffman M, Bratti C, Hildesheim A, et al. High load for most high risk human papillomavirus genotypes is associated with prevalent cervical cancer precursors but only HPV16 load predicts the development of incident disease. Int J Cancer 2007;121: 2787–93. [46] Huang Y, Huang MN, Li N, Li XG, Wu LY. Association between human papillomavirus DNA load and development of cervical intraepithelial neoplasia and cervical cancer. Int J Gynecol Cancer 2008;18:755–60. [47] Kimple AJ, Torres AD, Yang RZ, Kimple RJ. HPV-associated head and neck cancer: molecular and nano-scale markers for prognosis and therapeutic stratification. Sensors 2012;12:5159–69. [48] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. [49] Lee PN, Forey BA, Coombs KJ. Systematic review with meta-analysis of the epidemiological evidence in the 1900 relating smoking to lung cancer. BMC Cancer 2012;12:385.
Please cite this article in press as: Zhai K, et al. HPV and lung cancer risk: A meta-analysis. J Clin Virol (2014), http://dx.doi.org/10.1016/j.jcv.2014.09.014
