Abstract
Objective
The human papillomavirus (HPV) is implicated in the pathogenesis of several cancers among humans. The role of HPV as one of the etiological agents in esophageal carcinogenesis is partially unknown. We assessed whether the available evidence supports the association of HPV with risk and prognosis in patients with esophageal squamous cell carcinomas (ESCCs).
Design
For this systematic review and meta-analysis, PubMed, Embase, Cochrane Library, and SCOPUS were searched up to February 2021. The included studies were prospective or retrospective studies that evaluated the incidence, risk, and prognosis of HPV-16/18–related ESCCs in adult subjects. The primary outcome was the incidence rate of ESCC in HPV-16/18 carriers. Secondary outcomes included the risk of ESCCs compared with healthy HPV-16/18 carriers (expressed as odds ratios [ORs] with 95% confidence intervals [CIs]) and the survival of HPV + versus HPV- ESCCs.
Results
The search identified 1649 unique citations, of which 145 met the inclusion criteria and were included in the pooled analysis (16,484 patients). The pooled HPV prevalence in ESCCs was 18.2% (95% CI 15.2–21.6%; P < 0.001). A significantly increased ESCC risk was associated with HPV infection (OR = 3.81; 95% CI 2.84–5.11; P < 0.001). Main limitation were methods of HPV detection (DNA only), race of populations included (mainly Asiatic countries) and lack of adjustment for other prognostic factors.
Conclusions
The findings suggest that HPV-16/18 is detectable in about 1 on 5 cases of ESCC with different prevalences across the world. It is moderately but significantly associated with a diagnosis of ESCC. Further epidemiological studies are needed to confirm and increase the current knowledge of the subject.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00432-021-03738-9.
Keywords: Human papillomavirus, Infection, Esophageal cancer, Squamous, Risk, Meta-analysis
Introduction
Squamous cell carcinoma of the esophagus is a fatal disease with a low cure rate. It is the 8th most common cancer globally and the 6th most common cause of cancer deaths, with wide variation in incidence around the world.
This disease's etiology remains unknown, but extensive epidemiological and experimental studies have suggested several risk factors, such as cigarette smoking; the intake of excess alcohol; the intake of hot foods; specific nutritional deficiencies, including vitamins and minerals; and some chemical substances in food, especially nitrosamines and their precursors (Kok et al. 1997).
Human papillomaviruses (HPVs) have been found to play a pathogenetic role in cervical dysplasia and carcinoma. Like anogenital and other aerodigestive cancers, sufficient data are now available to implicate HPVs as one of the potential etiological agents in esophageal carcinogenesis (Kamath et al. 2000).
Several data suggest that some types of HPVs are the necessary factors of cervical oncogenesis, notably HPV-16, which can be identified in most invasive cancers and in their direct precursors (Tsikouras et al. 2016).
The data for esophageal cancers are conflicting. The geographic variation of the incidence of esophageal cancer is relevant. For example, when global incidence rates are examined, the incidence of esophageal cancer is 20-fold greater in high-risk China than in low-risk western Africa. Other areas with a relatively high risk of cancer include southern and eastern Africa, south-central Asia, Japan, and Korea. The association of HPVs with esophageal carcinogenesis seems to be significantly greater in high-risk areas compared to low-risk regions (Koh et al. 2008).
Therefore, the purpose of this study was to investigate the detection rates, the associated risk, and the prognosis of HPV-related ESCC in different areas of the world.
Material and methods
This meta-analysis was conducted according to the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) guidelines (Stroup et al. 2008).
Literature search
PubMed, Embase, Cochrane Library, and Web of Science databases were searched up to February 2021. PubMed, Embase, the Cochrane Library, and SCOPUS were searched with the following search terms: "esophageal neoplasms" (MeSH Terms) OR ("esophagus" [All Fields] AND "neoplasms" [All Fields]) OR "esophageal neoplasms" (All Fields) OR ("esophagus" [All Fields] AND "cancer" [All Fields]) OR "EC" (All Fields) OR ("ESCC" [All Fields] AND "human papillomavirus" or "HPV" [All Fileds]).
Inclusion criteria
The studies that evaluated HPV-16/18 viruses' positivity in cases with at least ten histologically proven ESCCs and eventually compared them with healthy controls were eligible. The patients with esophageal adenocarcinoma were ineligible. A cohort study, a cross-sectional study, a prospective series, or a case–control study was considered. The detection of viruses was performed in cases and controls based on tissue specimens. Any other genotype other than 16 and 18 was not eligible for this systematic review. No limitation was placed on race, sex, age, cancer stage, and treatment strategy. Only those studies with extractable data were considered for analysis. For an overlap study, only the most recent study or the study with the more significant number of participants was included.
Data extraction and quality assessment
Three reviewers (AG, AC, and FP) screened the titles/abstracts separately and assessed the potentially eligible full texts according to the predefined inclusion and exclusion criteria. Discrepancies were resolved by consensus with a fourth reviewer (URF). Three reviewers independently performed the risk of bias assessment of all included studies according to the Newcastle–Ottawa Scale (NOS) (Wells et al. 2012).
Nine reviewers extracted the study data independently. The information included general study characteristics (publication year, country, and study design), sample size, and types of measurements (specimen types and viruses detection methods). The numbers of cases and controls were extracted from publications. For a study that had paired corresponding normal tissues from the same patients with ESCC, as well as noncancer mucosa samples from patients without cancer as controls simultaneously, only data on patients without cancer were extracted. Potential discrepancies in the extracted items were resolved by further review and discussion by consensus with senior authors (MV and AL).
Statistical analysis
The pooled prevalence of the 16/18 types of virus infection in patients with EC was combined for the rate, with the confidence intervals (CIs) being 95%. Sensitivity analyses were performed according to the type of study, the year (published in the past 10 years vs. more than 10 years ago), the quality of the paper, and the country. According to heterogeneity, the pooled ORs and their 95% CIs for virus prevalence were calculated between patients with EC and any controls by the fixed or random effects model. After potential influencing factors were identified, multivariate logistic regression analysis was conducted in a stepwise fashion to further investigate factors that contribute to risk association. Two-sided P values for the pooled ORs < 0.05 were considered to be statistical significant. I2 was estimated to evaluate the heterogeneity. Any P value < 0.05 of the Begg or Egger test was considered to represent the significance of publication bias. The leave-one-out method was applied for sensitivity analysis. For prognostic meta-analysis, the aggregated hazard ratio (HR) and their 95% CIs for overall survival (OS) were calculated between EC patients with HPV + and HPV- via the random effects model due to the included studies' observational nature.
The comprehensive meta-analysis software and RevMan software were used for statistical analysis.
Results
A total of 2891 citations were identified. Overall, 145 studies were eligible for inclusion in the present meta-analysis (Fig. 1; Table 1 and suppl. File 1). A total of 16,484 participants were evaluated between 1992 and 2021.
Fig. 1.
Flow diagram of included studies
Table 1.
Characteristics of included studies
| Author/year | Type of study | Country | N° ESCC pts (case/controls) | HPV16/18 + (n) | HPV test | Tissue | Risk of cancer (OR, 95% CI) | Median follow up (months) |
Prognosis of HPV + cases (OS: HR, 95% CI) | NOS |
|---|---|---|---|---|---|---|---|---|---|---|
| Abdirad/2012 | Retrospective | Iran | 93 | 3/1 | PCR | FFPE | – | – | – | 5 |
| Akutsu/1995 | Retrospective | Japan | 29 | – | PCR | FFPE | – | – | – | 5 |
| Antonsson/2010 | Case–control | Australia | 222/55 | 6/0 | PCR | FFPE | 3.33 (0.18–60.05) | – | – | 6 |
| Antunes/2013 | Case–control | Brazil | 52/37 | – | PCR | FFPE | NE | – | – | 5 |
| Astori/2001 | Retrospective | Italy | 17 | 4/0 | PCR | FF | – | – | – | 5 |
| Awerkiew/2003 | Retrospective | Germany | 23 | – | PCR | FF | – | – | – | 5 |
| Bahnassy/2005 | Retrospective | Egypt | 47 | 12/6 | PCR | FFPE | – | 36 | – | 7 |
| Bas/2019 | Retrospective | Egypt/Somalia | 92 | 1/1 | PCR | FFPE | – | – | – | 5 |
| Benamouzig/1992 | Retrospective | France | 75 | – | PCR | FFPE | – | – | – | 5 |
| Bjorge/1997 | Case–control | Norway | 41/123 | 4/6 | Serum Ab | – | 6.08 (2.32–15.94) | – | – | 5 |
| Bognar/2018 | Retrospective | Hungary | 74 | 14/0 | ISH | FFPE | – | – | – | 5 |
| Cao B/2005 | Case–control | China | 265/357 | 175/4/7 both | PCR | FFPE | 2.09 (1.50–2.93) | – | – | 6 |
| Cao F/2014 | Retrospective | China | 105 | 25/0 | ISH | FFPE | – | 60 | – | 8 |
| Castillo/2006 | Retrospective | Various | 73 | 11/10 | PCR | FFPE | – | – | – | 5 |
| Castillo/2011 | Retrospective | Various | 166 | 24/0 | PCR | FFPE | – | – | – | 5 |
| Chang/1992 | Retrospective | China | 71 | 9/8 | PCR/ISH |
FFPE (51) FF (20) |
– | – | – | 5 |
| Chang/1993 | Retrospective | Finland/China | 363 | 16/7 | ISH | FFPE | – | – | – | 6 |
| Chang/1994 | Retrospective | Finland | 21 | 3/1 | ISH | FFPE | – | – | – | 5 |
| Chang/2000 | Retrospective | Various | 700 | 21/13 | ISH | FFPE | – | – | – | 6 |
| Cheah/2018 | Retrospective | Malaysia | 51 | 1/0 | PCR | FFPE | – | – | 5 | |
| Chen/1994 | Retrospective | US/China | 40 | 6/0 | PCR | FFPE | – | – | – | 5 |
| Chen/2014 | Matched case–control | China | 66/66 | – | PCR | FFPE | 14.50 (5.90–35.63) | – | – | 5 |
| Cooper/1994 | Retrospective | South Africa | 48 | 21/3 | NISH | FFPE | – | – | – | 5 |
| Dąbrowski/2012 | Case–control | Poland | 56/35 | – | PCR | Tissue samples | 7.75 (2.42–24.86) | – | – | 5 |
| DaCosta/2017 | Prospective cohort | Brazil | 123 | – | PCR | FFPE | – | – | – | 6 |
| Dai/2007 | Case-only | Various | 100 | 10/0 | PCR | FF | – | – | – | 6 |
| DeVilliers/1999 | Retrospective | Various | 117 | – | PCR | FF | – | – | – | 6 |
| Ding/2010 | Retrospective | China | 17 | 8/0 | PCR | FFPE | – | – | – | 5 |
| Dreilich/2006 | Retrospective | Sweden | 100 | 12/0 | PCR | FFPE | – | 10 | 0.71 (0.39–1.29) | 6 |
| Far/2007 | Retrospective | Iran | 140 | 33/0 | PCR | FFPE | – | – | – | 5 |
| Farhadi/2005 | Case–control | Iran | 38/38 | 8/0 | PCR | FFPE | 1.76 (0.52–5.97) | – | – | 5 |
| Fidalgo/1995 | Case–control | Portugal | 16/10 | 9/0 | PCR | FFPE | 26.60 (1.33–531.19) | – | – | 5 |
| Fukahori/2021 | Retrospective | Japan | 33 | 1/0 | PCR/ISH | FFPE | – | – | – | 5 |
| Furihata/1993 | Retrospective | Japan | 71 | 24/0 | PCR | FFPE | – | – | 1.46 (0.73–2.92) | 5 |
| Gao/2006 | Case–control | China | 42/660 | 42/0 | PCR | FFPE | 1.30 (0.53–3.19) | – | – | 5 |
| Georgantis/2015 | Case–control | Greece | 19/30 | 2/0 | PCR | FFPE | 8.71 (0.40–192.04) | – | – | 5 |
| Geßner/2018 | Case–control | Malawi | 40/12 | 6/0 | PCR/ISH/p16 | FFPE | 4.71 (0.25–89.84) | – | – | 5 |
| Goepfert/2009 | Retrospective | Mexico | 60 | 15/0 | PCR | FFPE | – | – | – | 5 |
| Goto/2011 | Retrospective | Asian Countries | 181 | 17/0 | PCR | FFPE | – | – | – | 6 |
| Guo/2012 | Case–control | China | 300/900 | 72/0 | PCR | FFPE | 10.21 (6.41–16.27) | – | – | 6 |
| Gupta/2012 | Case–control | India | 44/17 | 17/0 | PCR | FFPE | 22.27 (1.26–394.52) | – | – | 5 |
| Haeri/2013 | Case–control | Iran | 30/30 | – | PCR | FFPE | NE | – | – | 5 |
| Han/1996 | Case–control | Japan | 90/121 | 22/0 | PCR | FFPE | 4.57 (1.93–10.83) | – | – | 5 |
| Hasegawa/ 2002 | Retrospective | Japan | 48 | 20/0 | PCR | FFPE | – | – | – | 5 |
| He/1997 | Retrospective | China | 152 | 37/0 | PCR | FFPE/FF | – | – | 1.30 (0.64–2.64) | 5 |
| Herbster/2012 | Retrospective | Brazil | 264 | 34/0 | PCR/ISH | FFPE | – | 60 | – | 6 |
| Hille/1986 | Pilot study | South Africa | 70 | 7/0 | Serum Ab | – | – | – | 5 | |
| Hu/2012 | Case–control | Northwest China | 200/150 | 82/0 | PCR | FFPE | 4.27 (2.49–7.33) | – | – | 6 |
| Hu/2013 | Retrospective | Northwest China | 60 | 21/0 | PCR | FFPE | – | – | – | 5 |
| Hussain/2009 | Case–control | India | 75/75 | 14/0 | PCR/Immunoblotting | FFPE | 35.60 (2.08–608.88) | - | - | 5 |
| Inoue/2020 | Retrospective | Japan | 145 | 15/0 | ISH | FFPE | – | 24 | – | 6 |
| Ishida/2020 | Retrospective | Japan | 82 | – | PCR | FFPE | – | 60 | – | 7 |
| Iyer/2010 | Case–control | America | 120/29 | 41/0 | PCR | FFPE | 1.24 (0.49–3.18) | – | – | 6 |
| Kamangar/2006 | Case–control | China | 99/381 | 12/8 | Serum Ab | – | 8.28 (3.91–17.53) | – | – | 6 |
| Kamath/2000 | Retrospective | America | 46 | 1/0 | PCR | FFPE | – | – | – | 5 |
| Kanaan/2020 | Retrospective | France | 86 | - | RISH | FFPE | – | 6 | – | 5 |
| Katiyar/2005 | Case–control | India | 101/26 | 17/2 | PCR | FFPE | 5.79 (0.74–45.46) | – | – | 5 |
| Kayamba/2015 | Case–control | Zambia | 44/48 | 3/0 | PCR | FFPE | 2.24 (0.20–25.58) | – | – | 5 |
| Khurshid /1998 | Case–control | Japan | 27/12 | 3/4 | PCR | FFPE | 5.10 (1.11–23.37) | – | – | 5 |
| Koh/2007 | Case–control | Korea | 102/40 | – | PCR | FFPE | NE | – | – | 5 |
| Kok/1997 | Retrospective | Amsterdam | 63 | – | PCR | FFPE | – | – | – | 5 |
| Koshiol/2010 | Retrospective | China | 272 | – | PCR | FFPE | – | – | – | 6 |
| Koshiol2010 | Retrospective | Australia | 222 | 8/0 | PCR | FFPE | – | – | – | 6 |
| Kumar/2016 | Retrospective | India | 101 | 22/0 | PCR | FFPE | – | – | – | 5 |
| Lagergren/2020 | Case–control | Sweden | 294/302 | – | Serum Ab | – | 0.77 (0.44–1.34) | – | – | 5 |
| Lam/1997 | Retrospective | China | 70 | 7/0 | PCR | FFPE | – | – | – | 5 |
| Lavergne/1999 | Retrospective | China/South Africa | 63 | 0/26 | PCR | FFPE | – | – | – | 5 |
| Lee/2020 | Retrospective | China | 64 | – | PCR | FFPE | – | – | – | 5 |
| Leon/2019 | Case–control | Ethiopia | 62 | 1/0 | PCR/Luminex | FF | – | – | – | 5 |
| Li/2017 | Observational | China | 189 | 76/0 | PCR | FF | – | – | – | 6 |
| Li/2018 | Observational | China | 189 | 0/33 | PCR | FF | – | – | – | 6 |
| Liu/2009 | Cohort | China | 69/32 | 35/0 | PCR | FFPE | 16.99 (3.78–76.37) | – | 0.64 (0.31–1.32) | 5 |
| Liu/2013 | Observational | China | 253 | 52/0/6 both | qRT-PCR | FFPE | – | – | – | 5 |
| Liu/2014 | Cohort | China | 78/67 | 54/12 | PCR | FFPE/FF | – | – | – | 5 |
| Liyanage/2014 | Case–control (pooled analysis) | Australia | 321/155 | 9/0 | PCR | FFPE | 9.45 (0.55–163.50) | – | – | 6 |
| Lofdahl/2012 | Observational | Sweden | 204 | 13/0 | PCR | FFPE | – | – | – | 6 |
| Loke/1990 | Observational | China | 37 | – | FISH | FFPE | – | – | – | 5 |
| Lu/2007 | Retrospective | China | 67 | 7/2 | PCR | FFPE | – | – | – | 5 |
| Ludmir/2014 | Observational | US | 37 | 1/0 | PCR | FFPE | – | 129.4 | – | 5 |
| Lyronis/2005 | Case–control retrospective | Greece | 30/27 | 1/15 | PCR | FFPE | 4.00 (1.26–12.72) | – | – | 5 |
| Lyronis/2008 | Case–control retrospective | Greece | 30/32 | 17/6 | PCR | FFPE | 5.67 (1.80–17.79) | – | – | 5 |
| Malik/2011 | Retrospective | US | 11 | – | ISH | FFPE | – | – | – | 5 |
| Matsha/2002 | Retrospective | South Africa | 50 | 2/0 | PCR | FFPE | – | – | – | 5 |
| Matsha/2007 | Cohort | South Africa | 114/41 | 6/0 | PCR | FFPE | 4.97 (0.27–90.24) | – | – | 5 |
| Mehryar/2015 | Observational | China | 198 | 80/94 | PCR | FFPE | – | – | – | 5 |
| Miller/1997 | Observational | Alaska | 32 | 7/0 | PCR | FFPE | – | – | – | 5 |
| Mir/2007 | Retrospective | India | 62 | – | PCR | FFPE | – | – | – | 5 |
| Mizobuchi/1997 | Retrospective | Japan | 41 | – | PCR | FFPE | – | – | – | 5 |
| Mohamed/2021 | Retrospective | Sudan | 40 | 14/0 | PCR | FFPE | – | – | – | 5 |
| Mohiuddin/2013 | Case–control | India | 56/59 | 11/27 | PCR | FF | 0.29 (0.13–0.67) | – | – | 5 |
| Morgan/1997 | Retrospective | UK | 22 | – | PCR | FFPE | – | – | – | 5 |
| Nakamura/1995 | Retrospective | Japan | 61 | 5/0 | IHC | FFPE | – | – | – | 5 |
| Noori/2012 | Case–control | Iran | 92/20 | – | PCR | FFPE | NE | – | – | 5 |
| Ono/1994 | Retrospective | Japan | 42 | 15/0 | FITC | FFPE | – | – | – | 5 |
| Pandilla/2013 | Retrospective | India | 60 | 3/0 | PCR | FFPE | – | – | – | 5 |
| Pantelis/2007 | Retrospective | Germany | 53 | 9/0 | PCR | FFPE | – | – | – | 5 |
| Pastrez/2017 | Case–control | Brazil | 87/87 | 3/1 | PCR | FFPE | 3.07 (0.31–30.12) | – | – | 5 |
| Patel/2011 | Retrospective | Kenya | 28 | – | PCR | FFPE | – | – | – | 5 |
| Peixoto/2001 | Case–control | China | 32/57 | 2/4 | PCR | Citology | 0.88 (0.15–5.11) | – | – | 5 |
| Poljak/1998 | Retrospective | Slovenia | 121 | – | PCR | FFPE | – | – | – | 5 |
| Prakash Saxena/2016 | Retrospective | India | 18 | 9/0 | Hybrid Capture 2 | FFPE | – | 13.2 | – | 6 |
| Qi Z/2013 | Case–control | China | 225/224 | 87/65 | ELISA | Plasma | 1.54 (1.04–2.29) | – | – | 6 |
| Qi ZL/2006 | Retrospective | China | 60 | 24/0 | ISH | FFPE | - | - | - | 5 |
| Rugge/1997 | Retrospective | Italy/Egypt/Maryland | 18 | – | PCR/NISH | FFPE | – | – | – | 5 |
| Saegusa/1997 | Retrospective | Japan | 103 | – | PCR | FFPE | – | – | – | 5 |
| Shen/2002 | Retrospective | China | 77 | 30/17 | PCR | FF | – | – | – | 5 |
| Shibagaki/1995 | Retrospective | Japan | 72 | 15/0 | PCR | FF | – | – | – | 5 |
| Shuyama/2007 | Retrospective | China | 59 | 19/0 | PCR | FFPE | – | – | – | 5 |
| Si/2002 | Retrospective | China | 319 | 39/6 | PCR |
FFPE (75) FF (244) |
– | – | – | 5 |
| Sitas/2007 | Case–control | South Africa | 369/2055 | 134/0 | Serum Ab-ELISA | – | 1.65 (1.31–2.08) | – | – | 6 |
| Sitas/2011 | Case–control | Australia | 1561/2502 | 351/624 | Serum Ab | – | – | – | – | 7 |
| Smits/1995 | Retrospective | Netherlands | 63 | – | PCR | FFPE | – | – | – | 6 |
| Soheili/2015 | Retrospective | Iran | 103 | 6/5 | PCR | FFPE | – | – | – | 6 |
| Souto Damin/2006 | Case–control | Brazil | 165/26 | 24/1/1 both | PCR | FFPE | 10.07 (0.60–170.35) | – | – | 6 |
| Sultana/2020 | Retrospective | Pakistan | 203 | 41/16/6 both | PCR | FFPE | – | – | – | 6 |
| Suzuk/1996 | Retrospective | China/Ohio | 110 | 3/0 | PCR | FFPE | – | – | – | 6 |
| Takahashi/1998 | Retrospective | Japan | 123 | 22/23/8 both | PCR/ISH/DBH | FFPE | – | – | – | 6 |
| Tasneem/2019 | Descriptive, cross sectional | Pakistan | 121 | 4/0 | Monoclonal mouse Ab anti HPV | FFPE | – | – | – | 6 |
| Teng/2014 | Retrospective | China | 117 | 2/0 | PCR | FFPE | – | – | – | 6 |
| Togawa/1994 | Retrospective | Massachusetts/South Africa/ Maryland | 72 | 9/1 | PCR |
FFPE (25) FF (47) |
– | – | – | 5 |
| Toh/1992 | Retrospective | Japan | 45 | 1/2 | PCR | FF | - | - | - | 5 |
| Tornesello/2009 | Retrospective | Italy | 36 | 1/0 | PCR | FFPE | – | – | – | 5 |
| Türkay/2015 | Retrospective | Turkey | 33 | – | PCR | FFPE | – | – | – | 5 |
| Turner/1997 | Retrospective | US/Canada | 51 | 1/0 | PCR | FFPE | – | – | – | 5 |
| Vaiphei/2012 | Prospective | India | 23 | 2/0 | PCR | FF | – | – | – | 5 |
| Veyer/2019 | Prospective | France | 55 | 27/0 | PCR | FFPE | – | – | – | 5 |
| Wang JD/2013 | Retrospective | China | 61 | 0/26 | PCR | FFPE | – | 3.2–25.3 | – | 6 |
| Wang WL/2015 | Prospective | China | 150 | 22/0 | PCR | FFPE | – | 40.7 | 0.41 (0.18–0.93) | 7 |
| Wang X/2010 | Retrospective | China | 435 | 135/22 | PCR | FF | – | – | – | 5 |
| White/2005 | Prospective | Kenya | 29 | – | PCR | FFPE | – | – | – | 5 |
| Xi/2015 | Case–control | China | 103/54 | 65/0 | Ab | FFPE | 5.99 (2.81–12.75) | 60 | 1.87 (1.10–3.18) | 8 |
| Yahyapour/2013 | Case–control | Iran | 49/128 | 31/5 | PCR | FFPE | – | – | – | 5 |
| Yahyapour/2018 | Case–control | Iran | 100/68 | 7/0 | PCR | FFPE | 0.78 (0.25–2.42) | – | – | 6 |
| Yang W/2008 | Case–control | China | 435/550 | 304/0 | PCR | FFPE | 2.85 (2.18–3.72) | – | – | 7 |
| Yang Y/2013 | Case–control | China | 307/311 | 167/0 | Serum Ab | Blood samples | 1.56 (1.13–2.14) | – | – | 7 |
| Yao/2007 | Case–control | China | 82/40 | 23/0 | IHC/ISH | FFPE | 31.99 (1.89–541.82) | – | – | 5 |
| Zhang C-J/2020 | Case–control | China | 73/20 | – | ISH | FFPE | 33.91 (1.98–581.91) | 60 | – | 7 |
| Zhang D/2010 | Case–control | China | 70/60 | 40/15 | PCR | FFPE | 16.50 (5.34–51.03) | – | – | 5 |
| Zhang D/2017 | Retrospective | China | 67/192 | 43/0 | PCR | FFPE | – | – | 0.63 (0.41–0.97) | 6 |
| Zhang D-H/2011 | Case–control | China | 70/100 | 28/0 | PCR | FFPE | 7.67 (3.22–18.23) | – | – | 6 |
| Zhang D-H/2014 | Case–control | China | 70/50 | 35/0 | PCR | FFPE | 142.08 (8.43–2396.17) | 37.9 | – | 7 |
| Zhang Q-Y/2010 | Retrospective | China | 106/100 | 61/0 | PCR | FFPE | 4.81 (2.61–8.85) | – | – | 6 |
| Zheng/2020 | Retrospective | China | 54/40 | 21/0 | PCR | FFPE | 7.85 (2.14–28.73) | – | – | 6 |
| Zhou X-B/2003 | Case–Control | China | 48/23 | 31/0 | PCR/ISH | FFPE | 3.42 (1.21–9.69) | – | – | 5 |
| Zhou Y/2007 | Retrospective | China | 161 | 97/0 | PCR | FFPE | – | – | – | 6 |
HPV human papilloma virus, ESCC esophageal squamous cell carcinoma, OR odds ratio, OS overall survival, HR hazard ratio, CI confidence interval, PCR polymerase chain reaction, qRT–PCR real-time quantitative PCR, FISH fluorescence in situ hybridization, ISH in situ hybridization, NISH non-isotopic in situ hybridization, RISH mRNA in situ hybridization, IHC immunohistochemistry, FITC fluorescein isothiocyanate, ELISA enzyme-linked immunosorbent assay, Ab antibodies, FFPE formalin fixed paraffin embedded specimens, FF fresh frozens specimens, NE not estimable, – not reported
Characteristics of included studies
Among the included studies, 84 came from Asia, 22 from Europe, 11 from Africa, 11 from North and South America, 4 from Australia, and 13 from mixed populations. Studies were retrospective in 79 cases, case control in 48, and prospective/cohort or observational in 18.
We compared various HPV detection techniques: 109 studies used polymeraase chain reaction techniques, 4 serum antibodies, 9 in situ hybridization (ISH), 1 immunohistochemistry (IHC)-ISH, and the remaining multiple analyses. Furthermore, where information was declared, 119 studies used paraffin-embedded tissues for the analysis, 11 fresh frozen tissue, and 1 a blood sample.
The mean NOS Score is 5.4 (median 5).
HPVs and ESCC
A total of 141 studies investigated the prevalence of HPV infection, including 16,514 patients with ESCCs and 4525 cases of HPV infections. The pooled HPV prevalence in ESCCs was 18.2% (95% CI 15.2–21.6%, random model; heterogeneity P < 0.001).
According to type of infection, HPV-16 cases were reported in n = 113 studies, including 2506 cases of HPV-16 infections in 13 217 ESCCs. The pooled HPV-16 prevalence in ESCCs was 13.4% (95% CI 11.1–16.1%, random model; heterogeneity P < 0.001). There were 1198 cases of HPV-18 infections in 13,177 ESCCs (n = 112 studies). The pooled HPV-18 prevalence in ESCCs was 3% (95% CI 2.2–4.1%, random model; heterogeneity P < 0.001).
According to the country of origin, significant differences exist between Asiatic countries and Europe, North and South America, Australia, and Africa for HPV-16/18 infections (26.4%, 10.4%, 8.1%, 11.9%, 7.8%, and 12.1%, respectively; P < 0.001 for difference). This difference was not significant in a separate analysis of HPV-16 and HPV-18 cases. The type of analysis found different prevalences with higher values for polymerase chain reaction (PCR) (17.5%) and the lowest for ISH (12.1%). Serum antibodies or mixed techniques were associated with higher detection rates. In case–control studies, the pooled incidence rate was 41.4%. In cohort/prospective and retrospective studies, this was 26.1% and 17.1%. The incidence rate was 30% in studies published during or after 2010; meanwhile, it was 23.3% in those published before 2010.
In control arm (patients with no cancer), HPV infection rate was 12.6% (95%CI 9.5–16.6%).
Risk of ESCC and HPV infection
A total of 48 case–control or comparative studies with 5288 ESCC cases and 8085 controls were included in the estimation of HPV-16/18 infection and ESCC risk. According to the heterogeneity test result (Q test, P for heterogeneity < 0.001; I2 = 81%), the random effects model was chosen to evaluate the pooled ORs. As shown in Fig. 2, a significantly increased ESCC risk was associated with HPV infection (ORs = 3.81; 95% CI 2.84–5.11; P < 0.001). Furthermore, the summary OR of ESCC was 4.04 (95% CI 2.81–5.82%) for Asian patients with HPV-16/18 infection compared with those with no HPV infection. Western and African patients with HPV-16/18 infection had a lower risk of ESCC development (OR = 3.24, 95% CI 1.97–5.36; P < 0.001). According to the quality of the trials, ORs did not differ according to the NOS Score (ORs: 3.85 vs. 3.63 vs. 3.97 for NOS Scores of 5, 6, and ≥ 7).
Fig. 2.
Pooled ORs of ESCC associated with HPV-16/18 infection (cases vs. controls)
In multivariate logistic regression analyses, the percentage of smoker patients with squamous cell carcinoma was identified as a statistically significant influencing factor for an increased odds of ESCC (beta coefficient 2.75; P = 0.03).
Prognosis of HPV-related ESCCs
Only 7 studies reported the prognosis of HPV + in ESCC patients compared with HPV- patients. The overall survival was similar in the two populations, showing that the HPV status was not associated with better or worst survival (HR = 0.9, 95% CI 0.56–1.35; P = 0.6; Fig. 3).
Fig. 3.
Pooled HRs of overall survival for HPV-16/18 positive vs. negative cancers
Publication bias
Sensitivity and publication bias were conducted for studies involving the prevalence and risk of HPV infection in ESCC populations. In the leave-one-out analysis, we found that the pooled prevalence was always consistent by excluding any single study (range of 18.1% to 18.8%), which statistically verified the reliability of this meta-analysis.
Because all P values for the Begg test and Egger test were < 0.05, this meant that evidence of publication bias was identified. For the risk of ESCC in case–control studies, the funnel plot did not show obvious bias (P Begg's test 0.03). Conversely, Egger's test was significant (P = 0.002). When the overall homogeneity and effect size were calculated by removing one study at a time, we confirmed the stability of the positive association.
Discussion
Papillomaviruses are double-stranded deoxyribonucleic acid (DNA) viruses that constitute the Papillomavirus genus of the Papillomaviridae family. These viruses are highly species specific and infect the human race only. They are implicated in cervical, anogenital, and oropharyngeal carcinomas in a different association of causality. The virus's replication cycle is integrally linked to epithelial differentiation, and the initial infection of the basal stem cell occurs as a result of microscopic breaks in the epithelium. Evidence linking HPVs to cervical carcinoma is extensive (Garland 2002). Most cervical cancer cases are attributable to HPV infection, with HPV-16 accounting for approximately 50% of cases and HPV-18 accounting for 20% (Sanjose et al. 2010). HPV-associated oropharyngeal cancers are primarily found in the oropharynx and in the base of the tongue and tonsil (D'Souza et al. 2007; Gillison et al. 2012). The prevalence of oropharyngeal infection among males was 11.5% for any HPV type and 7.3% for high-risk HPV types; among the females, the prevalence was 3.2% and 1.4%, respectively (Sonawane et al. 2011).
HPV infections may be implicated in the pathogenesis of ESCC (Stoner and Gupta 2001; Eslick 2010; El-Zimaity et al. 2018). HPV-16/18 are the most studied types today, but a significant association has been deemed inconclusive, with at best a minority of ESCCs being related to HPVs. The previous meta-analyses were dated more than 5 years ago, describing different prevalences of HPV subtypes according to the type of analysis and country (Petrick et al. 2014; Li et al. 2014). Asiatic countries, as expected, are, in fact, geographical areas with the largest prevalences. HPV-16 is the genotype that is largely prevalent compared with HPV-18. We found in a mixed population of observational and case–control studies that high-risk HPV-16/18 infections are more prevalent in the Eastern part of the world (about double the prevalence of Europe, the Americas, and Africa), and a moderate link with ESCCs exists compared with healthy controls (about fourfold). Considerable variations exist across studies. However, association in case–control studies varies less significantly across the world (Asiatic vs. non-Asiatic subjects). We found in a multivariate regression analysis that the odds of ESCC risk were statistically correlated with smoking status but not sex, Asian race, or alcohol user. This is opposite of what has been found in head and neck cancers (Gillison et al. 2008). Detection of p16INK4a is the hallmark of potential HPV-related cancer (at least in oropharyngeal carcinomas), and the diagnosis is more accurately confirmed with DNA test by immunohistochemistry (IHC) or PCR. Despite near all studies used DNA tests, lack of p16 evaluation cannot be considered the gold standard, and so the definite correlation of HPV-16/18 with ESCC may not be established in this meta-analysis.
The genomic characterization of esophageal carcinoma published in Nature in 2017 has shown the existence of three subtypes of ESCCs through a molecular subclassification (Cancer and Atlas 2017). ESCC1 was characterized by alterations in the Nrf2 pathway, ESCC2 showed higher rates of the mutation of NOTCH1 or ZNF750, and ESCC3 tumors showed no evidence of the genetic deregulation of the cell cycle and had TP53 mutations in only 1 of 4 samples. HPV-related ESCC does not have a definite correlation with p53 expression; p53 mutations, in fact, are detectable in both HPV-positive and HPV-negative ESCCs, suggesting that HPVs and p53 mutations are not mutually exclusive events. The significant role of environmental carcinogens in ESCCs may be the case (Chang et al. 1994).
The association of HPVs with ESCC does not seem to be as strong as that observed for cervical and oropharyngeal cancers, and it presents a poor association with p16 as a biomarker of a high-risk infection (Michaelsen et al. 2014). An inferior association may be linked to anatomical and environmental reasons. First, high-risk sexual behaviors may not influence the esophageal epithelium and squamous cell cancer (more than adenocarcinoma) (Wong et al. 2018). Second, dietary factors (poor nutritional status, the low intake of fruits and vegetables, and drinking beverages at high temperatures) may be deemed risk agents in Asiatic countries. Conversely. Smoking status and alcohol abuse are major risk factors in Western countries (Tran et al. 2005; Li et al. 2011; Engel et al. 2003; Freedman et al. 2007). However, HPV + status can increase (as a synergistic carcinogenic factor) the promotion of ESCC, particularly in smokers and alcohol-user patients, as Qi et al. demonstrated, at least in China (2013). This may not occur in Western countries, as Georgantis et al. reported in a Greek population (2015). We also found no association of HPV (high-risk) status with survival in a few studies that reported outcome data. The association of HPV in Barrett's metaplasia and adenocarcinoma is also inconclusive, and no formal association can be discerned as observed in a systematic review and meta-analysis by Rajendra et al. in 2020.
Several limitations are indeed associated with the present analysis. First, most of the studies (about 50%) came from an Asiatic population with different genetic variables, ambiental causes, and lifestyles. Second, variations in the testing methodologies used, assay sensitivities, types of specimens utilized, and the years of analysis may have largely influenced the final results. It is expected, for example, that paraffin-embedded specimens can be subject to significant DNA degradation compared with fresh frozen samples. Despite IHC for p16INK4a being highly sensitive and the reference test to diagnose HPV associated tumors, most of the included studies confirmed the presence of high-risk HPV only with PCR. Third, the analysis of the risk association in case–control studies was made with crude data, and the weight of other prognostic variables (e.g., smoking status) was performed in only a few studies. Finally, only HPV-16/18 and only squamous histology were the focus of this meta-analysis.
A previous meta-analysis of 68 studies that Yong et al. published in 2013 found a positive association of HPVs in ESCC patients in Asiatic literature but not in non-Asiatic literature (Yong et al. 2013). Our meta-analysis, however, outperforms the past analysis because it includes 146 studies published up to a decade after the search limit of the Yong et al. publication. Also, we found a moderate and significant association of HPV + status with ESCCs in Asiatic and Western cohorts despite different geographical prevalences. In particular, HPV infection seems to play an additive role or even synergize with smoke. Although further epidemiological studies are needed to confirm this association's strength, this information is helpful for epidemiological and preventive/therapeutic purposes. Furthermore, appropriate studies adhering to methodological standards already established in other cancer entities should be used in ESCC to provide sufficient information about the pathogenesis of HPV-driven ESCC. Also, how different responses to treatment exist and how effective vaccines are in the prevention of HPV-related ESCC can be explored in future investigations.
Supplementary Information
Below is the link to the electronic supplementary material.
Declarations
Conflict of interest
None.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- Chang F, Syrjänen S, Tervahauta A, Kurvinen K, Wang L, Syrjänen S (1994) Frequent mutations of p53 gene in oesophageal squamous cell carcinomas with and without human papillomavirus (HPV) involvement suggest the dominant role of environmental carcinogens in oesophageal carcinogenesis. Br J Cancer 70(2):346–351. 10.1038/bjc.1994.305 [DOI] [PMC free article] [PubMed] [Google Scholar]
- de Sanjose S, Quint WGV, Alemany L et al (2010) Human papillomavirus genotype attribution in invasive cervical cancer: a retrospective cross-sectional worldwide study. Lancet Oncol. 10.1016/S1470-2045(10)70230-8 [DOI] [PubMed] [Google Scholar]
- D’Souza G, Kreimer AR, Viscidi R et al (2007) Case-control study of human papillomavirus and oropharyngeal cancer. N Engl J Med. 10.1056/nejmoa065497 [DOI] [PubMed] [Google Scholar]
- El-Zimaity H, Di Pilato V, Ringressi MN et al (2018) Risk factors for esophageal cancer: emphasis on infectious agents. Ann NY Acad Sci. 10.1111/nyas.13858 [DOI] [PubMed] [Google Scholar]
- Engel LS, Chow WH, Vaughan TL et al (2003) Population attributable risks of esophageal and gastric cancers. J Natl Cancer Inst. 10.1093/jnci/djg047 [DOI] [PubMed] [Google Scholar]
- Eslick GD (2010) Infectious causes of esophageal cancer. Infect Dis Clin North Am. 10.1016/j.idc.2010.08.001 [DOI] [PubMed] [Google Scholar]
- Freedman ND, Abnet CC, Leitzmann MF et al (2007) A prospective study of tobacco, alcohol, and the risk of esophageal and gastric cancer subtypes. Am J Epidemiol. 10.1093/aje/kwm051 [DOI] [PubMed] [Google Scholar]
- Garland SM (2002) Human papillomavirus update with a particular focus on cervical disease. Pathology. 10.1080/00313020212469 [DOI] [PubMed] [Google Scholar]
- Georgantis G, Syrakos T, Agorastos T et al (2015) Detection of human papillomavirus DNA in esophageal carcinoma in Greece. World J Gastroenterol. 10.3748/wjg.v21.i8.2352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gillison ML, D’Souza G, Westra W et al (2008) Distinct risk factor profiles for human papillomavirus type 16-positive and human papillomavirus type 16-negative head and neck cancers. J Natl Cancer Inst. 10.1093/jnci/djn025 [DOI] [PubMed] [Google Scholar]
- Gillison ML, Broutian T, Pickard RKL et al (2012) Prevalence of oral HPV infection in the United States, 2009–2010. JAMA. 10.1001/jama.2012.101 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Integrated genomic characterization of oesophageal carcinoma. Cancer Genome Atlas Research Network Nature 541(7636):169–175. 10.1038/nature20805. Accessed 12 Jan 2017 [DOI] [PMC free article] [PubMed]
- Kamath AM, Wu TT, Heitmiller R, Daniel R, Shah KV (2000) Investigation of the association of esophageal carcinoma with human papillomaviruses. Dis Esophagus. 10.1046/j.1442-2050.2000.00098.x [DOI] [PubMed] [Google Scholar]
- Koh JS, Lee SS, Baek HJ, Kim YI (2008) No association of high-risk human papillomavirus with esophageal squamous cell carcinomas among Koreans, as determined by polymerase chain reaction. Dis Esophagus. 10.1111/j.1442-2050.2007.00726.x [DOI] [PubMed] [Google Scholar]
- Kok TC, Nooter K, Tjong-A-Hung SP, Smits HL, Ter Schegget J (1997) No evidence of known types of human papillomavirus in squamous cell cancer of the oesophagus in a low-risk area. Eur J Cancer. 10.1016/S0959-8049(97)85984-9 [DOI] [PubMed] [Google Scholar]
- Li Y, Yang H, Cao J (2011) Association between alcohol consumption and cancers in the chinese population-a systematic review and meta-analysis. PLoS ONE. 10.1371/journal.pone.0018776 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Li X, Gao C, Yang Y et al (2014) Systematic review with meta-analysis: the association between human papillomavirus infection and oesophageal cancer. Aliment Pharmacol Ther. 10.1111/apt.12574 [DOI] [PubMed] [Google Scholar]
- Michaelsen SH, Larsen CG, Von Buchwald C (2014) Human papillomavirus shows highly variable prevalence in esophageal squamous cell carcinoma and no significant correlation to p16INK4a overexpression: a systematic review. J Thorac Oncol. 10.1097/JTO.0000000000000166 [DOI] [PubMed] [Google Scholar]
- Petrick JL, Wyss AB, Butler AM et al (2014) Prevalence of human papillomavirus among oesophageal squamous cell carcinoma cases: systematic review and meta-analysis. Br J Cancer. 10.1038/bjc.2014.96 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Qi Z, Jiang Q, Yang J et al (2013) Human papillomavirus (HPV) infection and the risk of esophageal squamous cell carcinoma. Dis Esophagus 26(1):61–67. 10.1111/j.1442-2050.2012.01334.x [DOI] [PubMed] [Google Scholar]
- Rajendra S, Pavey D, McKay O, Merrett N, Gautam SD (2020) Human papillomavirus infection in esophageal squamous cell carcinoma and esophageal adenocarcinoma: a concise review. Ann N Y Acad Sci. 10.1111/nyas.14509 [DOI] [PubMed] [Google Scholar]
- Sonawane K, Suk R, Chiao EY et al (2017) Oral human papillomavirus infection: Differences in prevalence between sexes and concordance with genital human papillomavirus infection, NHANES 2011 to 2014. Ann Intern Med. 10.7326/M17-1363 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stoner GD, Gupta A (2001) Etiology and chemoprevention of esophageal squamous cell carcinoma. Carcinogenesis. 10.1093/carcin/22.11.1737 [DOI] [PubMed] [Google Scholar]
- Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD (2008) MOOSE Guidelines for Meta-Analyses and Systematic Reviews of Observational Studies. JAMA 283:2008 [DOI] [PubMed] [Google Scholar]
- Tran GD, Di SX, Abnet CC et al (2005) Prospective study of risk factors for esophageal and gastric cancers in the Linxian General Population Trial cohort in China. Int J Cancer. 10.1002/ijc.20616 [DOI] [PubMed] [Google Scholar]
- Tsikouras P, Zervoudis S, Manav B et al (2016) Cervical cancer: screening, diagnosis and staging. J Buon 21:320 [PubMed] [Google Scholar]
- Wells G, Shea B, O'Connell D et al (2012) The Newcastle-Ottawa Scale (NOS) for assessing the quality if nonrandomized studies in meta-analyses.10.2307/632432
- Wong MYW, Wang B, Yang A, Khor A, Xuan W, Rajendra S (2018) Human papillomavirus exposure and sexual behavior are significant risk factors for Barrett’s dysplasia/esophageal adenocarcinoma. Dis Esophagus 31(12):1–7.10.1093/dote/doy051 [DOI] [PubMed] [Google Scholar]
- Yong F, Xudong N, Lijie T (2013) Human papillomavirus types 16 and 18 in esophagus squamous cell carcinoma: a meta-analysis. Ann Epidemiol.10.1016/j.annepidem.2013.07.002 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.