- Email: [email protected]
Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS Digestive and Liver Disease xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Digestive and Liver Disease journal homepage: www.elsevier.com/locate/dld
Meta-Analysis
Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis Edith Lahner ∗ , Marina Capasso, Marilia Carabotti, Bruno Annibale Medical-Surgical Department of Clinical Sciences and Translational Medicine, Sapienza University of Rome, Italy
a r t i c l e
i n f o
Article history: Received 11 January 2018 Received in revised form 16 May 2018 Accepted 17 May 2018 Available online xxx Keywords: Autoimmune gastritis Biliary tract cancer Cancer risk Leukaemia Lymphoma Meta-analysis Pernicious anaemia Systematic review
a b s t r a c t
Background: Pernicious anaemia (PA) is associated with increased gastric cancer risk, but the evidence is conflicting regarding the associated risk of other cancers. Aim: To systematically determine the incidence rates of gastro-intestinal cancers other than gastric cancers (GI-other-than-GC) and non-gastrointestinal cancers (non-GIC) in PA adults, globally and per tumour site, and the risk associated with PA for GI-other than GC and non-GIC. Methods: Studies of PA patients reporting the incidence of GI-other-than-GCs and non-GICs were identified with MEDLINE (PubMed)-EMBASE (from first date available to April 2017). A meta-analysis of annual cancer incidence rates was performed. The outcome was the cumulative incidence of GI-other-than-GCs and non-GICs (ratio between the numbers of new cancer cases identified during the follow-up period and the number of PA patients) and the incidence rate expressed as the rate of events-per-time-unit (person-years). Results: Of 82,257 PA patients, the pooled incidence rates/100 person-years for non-GCs and non-GICs of 0.27 (95% CI:0.16–0.42) and 0.23 (95% CI:0.22–0.25) were calculated by meta-analysis. Compared to the GLOBOCAN data for the general population from the countries of the included studies, the meta-analysis showed an overall relative risk (RR) of cancer in PA of 0.68 (95% CI:0.48–0.95). PA patients had a lower RR of colorectal, breast, liver, oesophageal, lung, thyroid, ovary, non-melanoma skin and kidney cancers but had a higher RR of biliary tract cancer (1.81:1.21–2.70), multiple myeloma (2.83:1.76–4.55), Hodgkin’s lymphoma (3.0:1.35–6.68), non-Hodgkin’s lymphoma (2.08: 1.58–2.75), and leukaemia (1.56:1.16–2.12). Conclusion: An overall lower RR of cancers-other-than-gastric-cancer in PA patients but an increased RR of biliary tract cancers and haematological malignancies was observed. © 2018 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
1. Introduction Pernicious anaemia (PA) is a macrocytic anaemia that results from cobalamin deficiency due to intrinsic factor deficiency, a protein avidly bound to dietary vitamin B12 that promotes its transport to the terminal ileum for absorption. Intrinsic factor deficiency is due to the presence of atrophic gastritis (AG), resulting in the loss of oxyntic mucosa and parietal cells and therefore the lack of chlorhydric acid as well as intrinsic factor secretion [1,2]. PA is considered a late stage of autoimmune gastritis, and it is often associated with positivity to parietal cell and/or intrinsic factor autoantibodies and
∗ Corresponding author at: Sapienza University of Rome, Department of MedicalSurgical Sciences and Translational Medicine, Via Grottarossa 1035, 00189 Rome, Italy. E-mail address: [email protected] (E. Lahner).
with the presence of other autoimmune disorders, usually autoimmune thyroid disease, type I diabetes, or vitiligo [3,4]. Similar to other autoimmune diseases, PA has been associated with an increased risk of cancer development [5]. A systematic review revealed a pooled gastric cancer incidence rate in PA patients of 0.3/100 person-years and an estimated 7-fold relative risk of gastric cancer [6]. In AG patients, type 1 gastric carcinoids may also occur [7–9]. High serum gastrin levels are found in PA as a response to damaged oxyntic mucosa and impaired gastric acid secretion. Hypergastrinaemia was shown to stimulate the growth of epithelial cells and to prevent apoptosis, possibly contributing to increased cancer risk [10]. In addition to gastric cancer and type 1 gastric carcinoids, PA has also been reported to be associated with an increased risk of cancers other than gastric neoplasias. Prior studies evaluating the possible association of PA with cancer risk have resulted in conflicting evidence; population-based studies mainly reported an increased odds ratio
https://doi.org/10.1016/j.dld.2018.05.012 1590-8658/© 2018 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
2
for haematological malignancies, such as monoclonal gammopathy of undetermined significance (MGUS) [11], multiple myeloma [12–14], acute myeloid leukaemia [14,15], chronic lymphocytic leukaemia [16], and myelodysplastic syndrome [14,15]. In contrast, a case-control study did not observe an association between PA and haematological malignancies [17]. Murphy et al. recently reported that patients with PA seem to be at increased risk for tonsillar cancer (OR 2.18), hypopharyngeal cancer (OR 1.92), oesophageal squamous cell carcinoma (OR 2.12), small intestinal cancer (OR 1.63) and liver cancer (OR 1.49) [14]. The aim of the present study was to determine the incidence rates of gastro-intestinal cancers other than gastric cancers (GIother than GC) and non-gastrointestinal cancers (non-GIC) in adult patients with PA, both globally and per tumour site, and to determine the risk associated with PA for GI-other than GC and non-GIC by systematically reviewing the literature. 2. Methods 2.1. Literature search strategy The literature on adult PA patients and the development of GI-other than GC and non-GIC was systematically reviewed. The search was conducted according to PRISMA guidelines [18]. The electronic databases MEDLINE (PubMed) and EMBASE were systematically searched combining the main keywords “pernicious anaemia”; “autoimmune gastritis”; and “cancer”. The search was translated into the following query: (“pernicious anaemia”[All Fields] OR “anaemia”, “pernicious”[MeSH Terms] OR (“anaemia”[All Fields] AND “pernicious”[All Fields]) OR (autoimmune[All Fields] AND (“gastritis”[MeSH Terms] OR “gastritis”[All Fields]) AND (“neoplasms”[All Fields] OR “cancer”[All Fields]) OR “carcinoma”[All Fields]) AND (“humans”[MeSH Terms] AND (English[lang] OR French[lang] OR German[lang] OR Spanish[lang]). No publication date restrictions were imposed. Reports published in English, German, French, Italian, and Spanish were considered. 2.2. Study selection Studies published from the first date available up to April 27th, 2017 were included in the systematic review if they fulfilled all of the following criteria: (1) an observational study including patients with PA and reporting the numbers of non-gastric and/or non-GI cancers identified during a defined follow-up period; (2) a study performed in adult patients; and (3) a study with an original full paper that presented unique data. Studies were excluded if (1) they were reviews, letters, editorials, or case-reports and (2) follow-up data were not available. Potentially relevant articles were screened for eligibility independently in an unblinded, standardized manner by the two reviewers (EL, MC), first by the abstract and then by the full text when necessary to determine whether they met the inclusion criteria. Disagreement between reviewers was resolved by discussion. The reference lists of the identified articles and relevant reviews were manually searched for additional studies that may have been missed using the electronic search strategy.
calculation of person-years (PY), age of patients (median or mean or range), gender, methods for identification of cancer, type of cancers, and numbers of cancer cases. 2.4. Outcome The outcome measures of interest were the cumulative incidence of GI-other than GCs and non-GICs, calculated as the ratio between the numbers of new cancer cases identified during the follow-up period and the number of PA patients, and the incidence rate was expressed as a rate of events per unit of time (personyears). 2.5. Statistical analyses From each included study, the number of incident non-GC and non-GIC cases and the number of PA patients exposed to risk were extracted. The cumulative incidence and the PY incidence rates in PA patients were calculated by the reviewers, if not explicitly stated. The cumulative incidence was calculated as the ratio between new non-GC and non-GIC cases and the number of PA patients. The PY incidence rates were calculated as the ratio between new non-GC and non-GIC cases and PY. The weighted summary proportion (pooled proportion) under the fixed and random effects model were calculated by a Freeman–Tukey transformation for all the non-GC and non-GIC cases and were calculated separately for specific cancer sites in single systems or organs. In the case of a positive “heterogeneity test” (p-value < 0.10), the more appropriate random effects model was taken into consideration, in which both the random variation within the studies and the variation between the different studies was incorporated [19]. The extent of heterogeneity was investigated using Cochran’s Q and I2 statistic [20]. The pooled incidence rates derived by meta-analysis were then compared with the annual cancer incidence rates of both genders in the general population aged over 40 years, as reported for the single European countries by the website GLOBOCAN (taking into consideration those countries to which the single studies included in the meta-analysis refer to) [21] by the Mantel–Haenszel method for calculating the weighted summary relative risk under the fixed effects model and the random effects model [19]. The statistical analysis was carried out using a dedicated software package (MedCalc Software, Mariakerke, Belgium, version 12.7.8.0). 2.6. Quality assessment The quality of all the included studies was evaluated based on the Newcastle-Ottawa quality assessment scale [22]. This scale awards a maximum of nine stars to each study. Three categories are considered: (i) cohort selection including four items, (ii) outcome assessment with three items, and (iii) comparability between cohorts and controls with two items. A study can be awarded a maximum of one star for each numbered item within the selection and outcome assessment categories and two stars for comparability. Studies were defined as of high quality when they obtained nine stars, of medium quality when they obtained seven or eight stars, of low quality when they obtained five or six stars, and of very low quality when they obtained four stars or less. Discrepancies in quality assessment were discussed and resolved by two reviewers (EL, MC).
2.3. Data extraction 3. Results Two reviewers (EL, MC) independently extracted the following information from each publication: name of the first author, publication year, country of study location, study design, criteria for diagnosis of PA, sources of participant selection, numbers of investigated patients, duration of follow-up period (expressed in years),
3.1. Search results The electronic search strategy identified a total of 2243 records from electronic databases and manual searches of the reference
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
3
Fig. 1. Flow chart of the study selection.
lists (Fig. 1). These articles were screened on the basis of the title and abstract. After application of the inclusion and exclusion criteria, 97 articles were retrieved for full-paper evaluation. Of these 97 fullpapers, 20 met the eligibility criteria and were subjected to data extraction [23–42]. 3.2. Quality assessment Details of the quality assessments of the included studies are given in Supplementary Table S1 in the online version at DOI: 10.1016/j.dld.2018.05.012. Of the twenty included studies, four (20.0%) [26,29,39,41] were of very low quality, five (25.0%) [23,24,28,30,33] of low quality, six (30.0%) [25,31,32,37,40,42] of medium quality and five (25.0%) [27,34–36,38] of high quality, according to the Newcastle-Ottawa quality assessment scale. 3.3. Characteristics of the included studies Female gender was prevalent in nine studies, with a range between 51.3% and 72.7%. In 1 study, there were fewer females than males (44.3%); in two studies, only male patients were investigated; and in the remaining studies, data on gender were lacking. The median age of patients was 71.4 years in all studies, ranging from 56.4 37–79.0 28 years.
The majority of the twenty studies were conducted in European countries, including nine (45.0%) in Northern Europe (Denmark, Finland, Sweden) and five (25.0%) in other European countries (United Kingdom, Germany). Five studies (25.0%) were American, and only one study (5.0%) was performed in China. Ten (50%) studies had a prospective design. The diagnostic criteria for PA diagnosis were thoroughly reported in ten (50.0%) studies. In nine studies, PA diagnosis was made by blood tests (macrocytic anaemia, low serum cobalamin or iron levels, hypergastrinaemia, low pepsinogen I levels), positive Schilling test, gastric pH, or megaloblastic bone marrow [23,24,26,28,33,37,39–41]; in seven studies, the data were extracted based on International Classification of Disease (ICD) [27,32,34–36,38,42]; and in one study, were based on hospital records [31]. In three studies [25,29,30], further details for the included PA patients were not provided. Most of the studies included patients who were first hospitalized for PA and afterwards were followed-up as outpatients. Three studies included outpatients referred to haematological units for anaemia or local healthcare providers [38–40]. In one study [25], patients were recorded in the files of eight PA clinics, but data were not provided on whether the patients were outpatients or hospitalized. Most of the studies retrieved the diagnosis of GI and non-GI cancers by biopsies using ICD codes from National Registers: three from local computer-based files [28,35,40] and four studies from
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
4
Table 1 Cancer incidence per organ system in pernicious anaemia (PA) patients. Cancer cases not included: 1 kidney/lymphoma/neurinoma [24]; 3 genital (not specified if female or male from 27); 24 not specified [26]; 26 not specified [27]; 44 not specified [25]; 1 metastasis [29]. Organ system
Cases of cancer/PA patients (cumulative incidence%) [annual incidence rate per 100.000 person-years]
Organ
Cases of cancer/PA patients (cumulative incidence%) [annual incidence rate per 100.000 person-years]
References
CNS
25/15,155 (0.2) [5.7]
Brain and meninges
25/15,155 (0.2) [5.7]
[24,27,28,34,36]
Maxillo-facial
206/30,042 (0.7) [8.4]
Sinus maxillary Eye Buccal cavity and pharynx
2/276 (0.7) [17.3] 1/1561 (0.06) [2.8] 203/30,004 (0.7) [9.1]
[24,29] [27] [27–29,32,34–36]
Respiratory
180/15,575 (1.2) [13.6]
Larynx Lung Other
2/4555 (0.04) [3.4] 165/15,537 (1.1) [13.6] 13/5161 (0.2) [37.0]
[24,34] [27–29,34–36,39] [27]
Endocrine
12/10,190 (0.1) [2.4]
Thyroid Other
5/5673 (0.1) [2.0] 7/4517 (0.1) [26.3]
[23,24,27,37,39] [34]
GI other than gastric
767/67,654 (1.1) [8.2]
Esophagus Small Bowel Colon Rectum Anus Liver Biliary tract Pancreas Other
141/50,902 (0.3) [3.6] 19/24,753 (0.08) [1.1] 271/30,789 (0.9) [10.3] 121/30,036 (0.4) [4.8] 3/11,839 (0.02) [4.4] 40/18,589 (0.2) [4.2] 30/15,478 (0.2) [6.7] 123/33,483 (0.4) [4.8] 19/5522 (0.4) [24.9]
[27,29,32,34–36,42] [27–29,32,35,36] [23,26,27,28,32–37,40] [24,26,27,29,32,34–36] [32] [26–28,34–37] [26–28,34,36] [26,27,29,34–36,38] [26,27]
Female reproductive
126/15,656 (0.8) [11.9]
Breast Uterus Ovary
79/15,656 (0.5) [7.5] 31/10,495 (0.3) [4.9] 16/9827 (0.2) [3.5]
[27–29,34,36,37] [28,29,34,36,37] [29,34,36]
Male reproductive
191/15,189 (1.3) [31.9]
Prostate
191/15,189 (1.3) [31.9]
[27,33,34,36,37]
Urinary
101/20,149 (0.5) [8.4]
Kidney Bladder
41/14,988 (0.3) [5.2] 60/14,988 (0.4) [7.6]
[26,29,34,36] [27,29,34,36]
Hematopoietic
199/29,713 (0.7) [10.0]
Hodgkin Lymphoma Non Hodgkin Lymphoma Multiple Myeloma Leukemia Other
3/9678 (0.03) [2.4] 70/22,240 (0.3) [15.4] 28/14,888 (0.2) [6.5] 76/16,557 (0.5) [10.8] 22/11,808 (0.2) [6.1]
[27,34] [31,34,36] [27,33–36] [25,27,30,34,36,39] [25,27,28,33,34]
Skin
86/14,750 (0.6) [32.8]
Melanoma Non melanoma
21/14,750 (0.1) [8.0] 65/9589 (0.7) [61.6]
[27,34,36] [34,36]
Bone and soft tissue
13/5766 (0.2) [4.9]
Bone Connective/soft tissue
6/5161 (0.1) [17.1] 7/5766 (0.1) [2.6]
[27] [27–29]
clinical interviews or death certificates [25,37,39,41]. Three studies did not specify how the cancer diagnoses were assessed [29,30,33]. For the data on cancer incidence, four studies reported only GI-other than GC cases [38,40–42], and three studies reported only non-GIC cases [30,31,39], while the remaining studies reported both cases. The characteristics of the included studies are summarized in Supplementary Tables S2 and S3 in the online version at DOI: 10.1016/j.dld.2018.05.012. 3.4. Cancer incidence With regard to the cancer incidence in PA patients, we found that the overall cumulative incidence was 2.4% for a cumulative follow-up of 185,5 years, ranging from 0.2% [42] to 15.8% [24], with 1993 new cancer cases in a total PA population of 82,257 patients (15,285,818.3 PY), corresponding to an overall incidence rate of 13.0/100,000 PY. In particular, the cumulative cancer incidence ranged from 0.2% [38,42] to 9.2% [34] in the 9 Northern European studies, from 0.2% [38] to 8.3% [41] in the 5 studies of other European countries, from 5.1% to 15.8% in the 5 American studies and 6.0% in the Chinese study. Overall, a total of 67,654 (9,392,404.8 PY) and 45,373 (6,825,006.7 PY) PA patients were studied, with 767 and 1194 new cancer cases in the GI-other than GC and non-GIC group, respectively. Examining GI-other than GCs, the cumulative incidence was 1.1% for a cumulative follow-up of 138.8 years (incidence rate
8.2/100,000-PY), ranging from 0.2% [38,42] to 8.0% [26] (Supplementary Table S2 in the online version at DOI: 10.1016/j.dld. 2018.05.012). Examining non-GICs, the cumulative incidence was 2.6% for a cumulative follow-up of 150.2 years (incidence rate 17.5/100,000-PY), ranging from 0.3% [32] to 13.2% [24] (Supplementary Table S3 in the online version at DOI: 10.1016/j.dld.2018. 05.012). The heterogeneity between studies was statistically significant in both groups (p < 0.0001). From the meta-analysis, a pooled incidence rate/100-PY for non-GCs and non-GICs of 0.27 (95% CI: 0.16–0.42) (Supplementary Fig. S1 in the online version at DOI: 10. 1016/j.dld.2018.05.012) and 0.23 (95% CI: 0.22–0.25) (Supplementary Fig. S2 in the online version at DOI: 10.1016/j.dld.2018.05.012) was found. Analysing the specific site of cancer incidence (Table 1), we found that the highest incidence rates/100,000-PY were observed for skin cancer (32.8, driven by non-melanoma cancer: 61.6) and prostate cancer (31.9). A meta-analysis performed using the annual cancer incidence rates per organ showed an overall relative risk (RR) for cancer in PA compared to the country-specific GLOBOCAN data of the general population (adults over 40 years of age) of 0.68 (95% CI: 0.48–0.95). With regard to site-specific cancers, PA patients had a lower RR (RR: 95% CI) compared to the general population for lung (0.26: 0.21–0.31), thyroid (0.34: 0.21–0.57), oesophagus (0.26: 0.18–0.37), colorectal (0.14: 0.01–0.19), liver (0.18: 0.13–0.24), pancreas (0.65: 0.45–0.93), breast (0.17: 0.13–0.22), ovary (0.45: 0.30–0.67), prostate (0.72: 0.63–0.84), kidney (0.61: 0.43–0.86),
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
5
Fig. 2. Relative risk of cancer per organ in patients with pernicious anaemia compared with the GLOBOCAN data for the general population (adult patients over 40 years of age) from the countries in which the included studies were performed. p < 0.0001; I2 = 96.79%; 95% CI: 95.92–97.47.
and non-melanoma skin cancers (0.56: 0.50–0.61). In contrast, PA patients had a higher risk than the general population for biliary tract cancer (1.81: 1.21–2.70) and for haematological malignancies, such as multiple myeloma (2.83: 1.76–4.55), Hodgkin’s lymphoma (3.0: 1.35–6.68), non-Hodgkin’s lymphoma (2.08: 1.58–2.75), and leukaemia (1.56: 1.16–2.12) (Fig. 2).
4. Discussion Evidence on overall cancer incidence in PA is sparse and conflicting. To our knowledge, this is the first systematic review of the estimate of GI-other than GCs and non-GICs cumulative incidence and incidence rates of PY in these patients. From 2243 articles retrieved from a search strategy with no time limit, we selected 20 papers according to our inclusion criteria published between 1950 and 2014, with a total of 82,257 PA patients (15,285,818.3 PY). Overall, 1993 new cases of cancer other than gastric cancer were identified, corresponding to a cumulative incidence of 2.4%, ranging from 0.2% to 15.8%. Furthermore, 767 and 1194 new cases of GI cancers other than GCs and non-GICs, respectively, were identified, with a cumulative incidence of 1.1% (incidence rate 8.2/100,000 PY) and 2.6% (15.7/100,000 PY), respectively. Compared with the GLOBOCAN data for the general population over 40 years of age (from the countries in which the studies were performed), PA patients had an overall lower RR (0.68) of developing cancers (other than gastric), even if we found an increased risk for haematological malignancies (in particular multiple myeloma [2.83] and Hodgkin’s lymphoma [3.0], and biliary tract cancers [1.8]). Subjects with PA have long been suggested to have an increased risk of gastric cancer as well as of gastric carcinoid tumours [6,8,9,30,43], but PA has been reported to be associated with an increased risk of other cancers. However, the paucity of previous studies on PA and cancer make this association controversial [10–15,17,35]. In our analysis, we found that PA patients had a lower RR of GI-other than GCs (oesophagus, liver, pancreas and colorec-
tum). The possible association between PA and GI-other than GCs has been debated in the literature, although the evidence is conflicting. An inverse relationship between gastric atrophy and oesophageal cancer has been observed in some studies [44–46], which is thought to reflect reduced acid reflux from the stomach to the oesophagus as a consequence of the oxyntic mucosa atrophy and the loss of parietal cells. In agreement with this finding, Murphy et al. found a positive association with oesophageal squamous cell carcinoma but no association with oesophageal adenocarcinoma [14]. As the authors stated, this relationship is not fully understood and may or may not be causal [14]. Gastrin and its precursors activate several pathways important in tumourigenesis, such as the beta catenin, MAP kinase and JAK2/STAT3 pathways, and they were described in association with colorectal, pancreatic and liver cancers in addition to gastric cancers and carcinoids [26,40,47]. Gastrin, at physiological concentrations, was shown to act as a growth factor for colorectal cancer through endocrine as well autocrine/paracrine mechanisms, and pre-malignant adenomas were also shown to express an isoform of the cholecystokinin B/gastrin receptor [48,49]. However, the lack of association between PA and colorectal cancer risk was recently demonstrated in a case control study [50] as well as in the largest population study currently available [10]. Only one study reported an increased risk of liver cancer for subjects with PA [34], which was recently confirmed in females [14]. No specific association of PA with lung, breast, prostate or kidney cancers was found in a recent population-based study [14], supporting our results. Thus, the findings of the present systematic review are in agreement with these data, denying the higher risk of cancers in these organs in patients with PA. However, in our analysis, we found an increased relative risk for haematological malignancies and biliary tract cancers. The role for immune dysregulation in haematological malignancies is well established, and several studies investigated a possible association between PA and haematological malignancies. Overall, PA seemed to increase the risk for multiple myeloma [12–14], acute myeloid
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7 6
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
leukaemia and myelodysplastic syndrome [14,15]. However, the association between PA and multiple myeloma was not confirmed by Lewis et al. [51] or Lindqvist et al. [11]. In addition, the study by Soderberg et al. [17] did not find any association between PA and all haematological diseases. The association between the risk of haematological malignancies and PA could be related to cobalamin deficiency [12,15]. Cobalamin deficiency may promote the development of cancer by causing derangement of one-carbon metabolism nutrients (folate, vitamin B6 , riboflavin, homocysteine) associated with aberrations in DNA methylation and DNA synthesis, as proposed by some authors [52,53]. When cobalamin is insufficient, cellular folate accumulates in the methylfolate form but cannot be metabolically utilized as a consequence of the block in the methionine synthase reaction, creating a functional folate deficiency referred to as the “methylfolate trap” [54]. Previous studies in bone marrow cells and immortalized cultured cells [55,56] demonstrated that cobalamin deficiency increased uracil misincorporation into DNA, suggesting that this functional folate deficiency by cobalamin deficiency results in inadequate thymidylate synthesis. Some authors observed that localized cobalamin deficiency in squamous-cell lung cancer tissue was associated with genomic DNA hypomethylation, suggesting that decreased availability of methylfolate by cobalamin deficiency subsequently reduces genomic DNA methylation [52]. We also found a positive association of PA with the risk of biliary tract cancers. There are no specific literature data on this association; however, a possible role for DNA hypomethylation can be hypothesized, considering that aberrant DNA methylation is an epigenetic mechanism that can occur early in carcinogenesis [57]. A recent molecular study demonstrated the up-regulation of Dicer in cholangiocarcinoma cells, promoting their proliferation and invasion [58]. Dicer is known to regulate methylation of CpG islands in mammalian cancer cells [59,60], interacting with heterochromatin protein 1␣ (HP1␣). DNA methylation is also considered one of the key molecular mechanisms in gallbladder tumourigenesis [61]. In a recent study, the aberrant methylation of the WIF-1 promoter was found to be involved in the malignant transformation of gallbladder cancer [62]. We are aware that this study has some limits. To evaluate the quality of included studies, we applied the Newcastle-Ottawa Quality Scale for cohort studies, which permits a critical appraisal of data [22]. We found that almost half of the included studies were of high/medium quality. In spite of this, clinical and methodological heterogeneity was present among the 20 studies included, resulting in a wide range of observed cumulative cancer incidence (0.2%–15.8%). The different approaches across the studies should take into account the decades in which the studies were performed. Diagnostic tools for both PA and cancer diagnosis have changed over time. A cancer diagnosis obtained from death certificates in the older studies could be not directly comparable with those obtained from ICD codes in National Cancer Registers. Additionally, six studies did not specifically report how the cancer was detected. Moreover, we excluded from this analysis the incidence rate for gastric cancer, as this issue was previously reported [6]. We know that the introduction of GI endoscopy has been a revolution in the diagnosis of upper GI diseases. This observation might impact the number of oesophageal cancers reported in older studies; some of those cases may include primary misdiagnosed gastric cancer. Notwithstanding these limits, we performed a meta-analysis based on the calculated GI-other than GCs and non-GICs incidence rates in PA patients. A further limitation of this systematic review is that for the comparison with the general population used for the calculations of the relative risk of cancer in PA, data from cancer registries (GLOBOCAN) were used, but the retrieved studies did not report control data, and this was a way to obtain estimates on the association of the single cancers with PA. To reduce this limitation, we used
as a comparison the GLOBOCAN data from adult patients over 40 years of age from those countries in which the studies included in the meta-analysis were performed. Although we are aware that the results of this meta-analysis may be biased by the limits explained above, in particular the structural fragility of the available information, we do think that they are acceptable as an estimate of the overall probability of PA patients to develop cancer and may represent a starting point to design further population-based prospective studies. In conclusion, this systematic review revealed an overall cancer risk (other than gastric cancer) for PA patients that was lower than the risk in the general population, except for haematological malignancies and biliary tract cancers. Further high-quality studies are needed to confirm the higher risk of these specific cancers in PA patients. Conflict of interest None declared. Funding This paper was funded in part by a grant from Sapienza University, funds of Oncology Doctorate 2016. Acknowledgments All the authors approved the final version of the article, including the authorship list. We are indebted to Mimma Ariano, Ales Casciaro, and Teresa Prioreschi of the University Library, Sapienza University, Sant’Andrea Hospital for their kind assistance during the retrieval of full text of papers. References [1] Annibale B, Lahner E, Delle Fave G. Diagnosis and management of pernicious anemia. Curr Gastroenterol Rep 2011;13:518–24. [2] Babior BM. Erythrocyte disorders: anemias related to disturbance of DNA synthesis (megaloblastic anemias). In: Williams JW, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 4th ed. New York: McGraw-Hill; 1998. p. 453–81. [3] Toh BH, Gleeson PA, Whittingham S, van Driel IR. Autoimmune gastritis and pernicious anemia. In: Rose NR, Mackay IR, editors. The autoimmune diseases. 3rd ed. Academic Press; 1998. p. 459–76. [4] Lahner E, Centanni M, Agnello G, Gargano L, Vannella L, Iannoni C, et al. Occurrence and risk factors for autoimmune thyroid disease in patients with atrophic body gastritis. Am J Med 2008;121:136–41. [5] Lahner E, Annibale B. Pernicious anemia: new insights from a gastroenterological point of view. World J Gastroenterol 2009;15:5121–8. [6] Vannella L, Lahner E, Osborn J, Annibale B. Systematic review: gastric cancer incidence in pernicious anaemia. Aliment Pharmacol Ther 2013;37:375–82. [7] Vannella L, Sbrozzi-Vanni A, Lahner E, Bordi C, Pilozzi E, Corleto VD, et al. Development of type I gastric carcinoid in patients with chronic atrophic gastritis. Aliment Pharmacol Ther 2011;33:1361–9. [8] Kokkola A, Sjöblom SM, Haapiainen R, Sipponen P, Puolakkainen P, Järvinen H. The risk of gastric carcinoma and carcinoid tumours in patients with pernicious anaemia. A prospective follow-up study. Scand J Gastroenterol 1998;33:88–92. [9] Lahner E, Esposito G, Pilozzi E, Purchiaroni F, Corleto VD, Di Giulio E, et al. Occurrence of gastric cancer and carcinoids in atrophic gastritis during prospective long-term follow-up. Scand J Gastroenterol 2015;50:856–65. [10] Boursi B, Mamtani R, Haynes K, Yang YX. Pernicious anemia and colorectal cancer risk — a nested case-control study. Dig Liver Dis 2016;48:1386–90. [11] Lindqvist EK, Goldin LR, Landgren O, Blimark C, Mellqvist UH, Turesson I, et al. Personal and family history of immune-related conditions increase the risk of plasma cell disorders: a population-based study. Blood 2011;118:6284–91. [12] Anderson LA, Gadalla S, Morton LM, Landgren O, Pfeiffer R, Warren JL, et al. Population-based study of autoimmune conditions and the risk of specific lymphoid malignancies. Int J Cancer 2009;125:398–405. [13] Brown LM, Gridley G, Check D, Landgren O. Risk of multiple myeloma and monoclonal gammopathy of undetermined significance among white and black male United States veterans with prior autoimmune, infectious, inflammatory, allergic disorders. Blood 2008;111:3388–94. [14] Murphy G, Dawsey SM, Engels EA, Ricker W, Parsons R, Etemadi A, et al. Cancer risk after pernicious anemia in the US elderly population. Clin Gastroenterol Hepatol 2015;13, 2282-9.e1-4.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
[15] Anderson LA, Pfeiffer R, Landgren O, Gadalla S, Berndt SI, Engels EA. Risk of myeloid malignancies in patients with autoimmune conditions. Br J Cancer 2009;100:822–8. [16] Landgren O, Engels EA, Caporaso NE, Gridley G, Mellemkjaer L, Hemminki K, et al. Patterns of autoimmunity and subsequent chronic lymphocytic leukemia in Nordic countries. Blood 2006;108:292–6. [17] Soderberg KC, Jonsson F, Winqvist O, Hagmar L, Feychting M. Autoimmune diseases, asthma and risk of haematological malignancies: a nationwide casecontrol study in Sweden. Eur J Cancer 2006;42:3028–33. [18] Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. [19] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88. [20] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. [21] GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. International Agency for Research on Cancer, World Health Organization. Available at: http://globocan.iarc.fr/old/factsheet.asp. [Accessed 24 July 2017]. [22] Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of non randomised studies in meta-analyses. Available at: http://www.ohri.ca/programs/clinical epidemiology/oxford.asp. [Accessed 24 July 2017]. [23] Armbrecht U, Stockbrügger RW, Rode J, Menon GG, Cotton PB. Development of gastric dysplasia in pernicious anaemia: a clinical and endoscopic follow up study of 80 patients. Gut 1990;31:1105–9. [24] Arvanitakis C, Holmes FF, Hearne 3rd E. A possible association of pernicious anemia with neoplasia. Oncology 1979;36:127–9. [25] Blackburn EK, Callender ST, Dacie JV, Doll R, Girdwood RH, Mollin DL, et al. Possible association between pernicious anaemia and leukaemia: a prospective study of 1625 patients with a note on the very high incidence of stomach cancer. Int J Cancer 1968;3:163–70. [26] Borch K, Kullman E, Hallhagen S, Ledin T, Ihse I. Increased incidence of pancreatic neoplasia in pernicious anemia. World J Surg 1988;12:866–70. [27] Brinton LA, Gridley G, Hrubec Z, Hoover R, Fraumeni JF. Cancer risk following pernicious anaemia. Br J Cancer 1989;59:810–3. [28] Chan JC, Liu HS, Kho BC, Lau TK, Li VL, Chan FH, et al. Longitudinal study of Chinese patients with pernicious anaemia. Postgrad Med J 2008;84:644–50. [29] Demmler K. Causes of death in patients treated for pernicious anemia. Med Klin 1966;61:575–7. [30] Elsborg L, Andersen D, Bastrup-Madsen P. Gastrocamera screening in pernicious anaemia. With special reference to the occurrence of gastric polyps and cancer. Scand J Gastroenterol 1973;8:5–8. [31] Fallah M, Liu X, Ji J, Försti A, Sundquist K, Hemminki K. Autoimmune diseases associated with non-Hodgkin lymphoma: a nation wide cohort study. Ann Oncol 2014;25:2025–30. [32] Hemminki K, Liu X, Ji J, Sundquist J, Sundquist K. Autoimmune disease and subsequent digestive tract cancer by histology. Ann Oncol 2012;23:927–33. [33] Hoffman NR. The relationship between pernicious anemia and cancer of the stomach. Geriatrics 1970;25:90–5. [34] Hsing AW, Hansson LE, McLaughlin JK, Nyren O, Blot WJ, Ekbom A, et al. Pernicious anemia and subsequent cancer. A population-based cohort study. Cancer 1993;71:745–50. [35] Landgren AM, Landgren O, Gridley G, Dores GM, LInet MS, Morton LM. Autoimmune disease and subsequent risk of developing alimentary tract cancers among 4.5 million US male veterans. Cancer 2011;117:1163–71. [36] Mellemkjaer L, Gridley G, Møller H, Hsing AW, Linet MS, Brinton LA, et al. Pernicious anaemia and cancer risk in Denmark. Br J Cancer 1996;73:998–1000. [37] Mosbech J, Videbaek A. Mortality from and risk of gastric carcinoma among patients with pernicious anaemia. Br Med J 1950;2:390–4. [38] Shah P, Rhim AD, Haynes K, Hwang WT, Yang YX. Diagnosis of pernicious anemia and the risk of pancreatic cancer. Pancreas 2014;43:422–6.
7
[39] Siurala M, Varis K, Wiljasalo M. Studies of patients with atrophic gastritis: a 10-15-year follow-up. Scand J Gastroenterol 1966;1:40–8. [40] Talley NJ, Chute CG, Larson DE, Epstein R, Lydick E, Melton J. Risk for colorectal adenocarcinoma in pernicious anemia. A population-based cohort study. Ann Intern Med 1989;111:738–42. [41] von Knorre G, Pechau KG. Late fate of patients with pernicious anemia. Z Gesamte Inn Med 1975;30:701–6. [42] Ye W, Nyren O. Risk of cancers of the oesophagus and stomach by histology or subsite in patients hospitalised for pernicious anaemia. Gut 2003;52:938–41. [43] Sjöblom SM, Sipponen P, Järvinen H. Gastroscopic follow up of pernicious anaemia patients. Gut 1993;34:28–32. [44] Cook MB, Dawsey SM, Diaw L, Blaser MJ, Perez-Perez GI, Abnet CC, et al. Serum pepsinogens and Helicobacter pylori in relation to the risk of esophageal squamous cell carcinoma in the alpha-tocopherol, beta-carotene cancer prevention study. Cancer Epidemiol Biomarkers Prev 2010;19:1966–75. [45] Islami F, Sheikhattari P, Ren JS, Kamangar F. Gastric atrophy and risk of oesophageal cancer and gastric cardia adenocarcinoma—a systematic review and meta-analysis. Ann Oncol 2011;22:754–60. [46] Kamangar F, Diaw L, Wei WQ, Abnet CC, Wang GQ, Roth MJ, et al. Serum pepsinogens and risk of esophageal squamous dysplasia. Int J Cancer 2009;124:456–60. [47] Karlson BM, Ekbom A, Wacholder S, McLaughllin JK, Hsing AW. Cancer of the upper gastrointestinal tract among patients with pernicious anemia: a casecohort study. Scand J Gastroenterol 2000;35:847–51. [48] Singh P, Dai B, Wu H, Owlia A. A role of autocrine and endocrine gastrin-like peptides in colonic carcinogenesis. Curr Opin Gastroenterol 2000;16:68–77. [49] Smith AM, Watson SA. Gastrin and gastrin receptor activation — an early event in the adenoma-carcinoma sequence. Gut 2000;47:820–4. [50] Lahner E, Sbrozzi-Vanni A, Vannella L, Corleto VD, Di Giulio E, Delle Fave G, et al. No higher risk for colorectal cancer in atrophic gastritis-related hypergastrinemia. Dig Liver Dis 2012;44:793–7. [51] Lewis DR, Pottern LM, Brown LM, Silverman DT, Hayes RB, Schoenberg JB, et al. Multiple myeloma among blacks and whites in the United States: the role of chronic antigenic stimulation. Cancer Causes Control 1994;5:529–39. [52] Friso S, Choi SW. The potential cocarcinogenic effect of vitamin B12 deficiency. Clin Chem Lab Med 2005;43:1158–63. [53] Miranti EH, Stolzenberg-Solomon R, Weinstein SJ. Low vitamin B12 increases risk of gastric cancer: a prospective study of one-carbon metabolism nutrients and risk of upper gastrointestinal tract cancer. Int J Cancer 2017;141:1120–9. [54] Herbert V, Das KC. Folic acid and vitamin B12. In: Shils ME, Olson JA, Shike M, editors. Modern nutrition in health and disease, vol. 1. Philadelphia: Lea & Febiger; 1994. p. 402–25. [55] Wickramasinghe SN, Fida S. Misincorporation of uracil into the DNA of folateand B12-deficient HL60 cells. Eur J Haematol 1993;50:127–32. [56] Wickramasinghe SN, Fida S. Bone marrow cells from vitamin B12- and folatedeficient patients misincorporate uracil into DNA. Blood 1994;83:1656–61. [57] Esteller M. Epigenetics in cancer. N Engl J Med 2008;358:1148–59. [58] Cheng W, Qi Y, Tian L, Wang B, Huang W, Chen Y. Dicer promotes tumorigenesis by translocating to nucleus to promote SFRP1 promoter methylation in cholangiocarcinoma cells. Cell Death Dis 2017;8:e2628. [59] Ting AH, Suzuki H, Cope L, Schuebel KE, Lee BH, Toyota M, et al. A requirement for DICER to maintain full promoter CpG island hypermethylation in human cancer cells. Cancer Res 2008;68:2570–5. [60] Chen Y, Luo J, Tian R, Sun H, Zou S. miR-373 negatively regulates methylCpG-binding domain protein 2 (MBD2) in hilar cholangiocarcinoma. Dig Dis Sci 2011;56:1693–701. [61] Tekcham DS, Tiwari PK. Epigenetic regulation in gallbladder cancer: promoter methylation profiling as emergent novel biomarkers. Asia Pac J Clin Oncol 2016;12:332–48. [62] Lin B, Hong H, Jiang X, Li C, Zhu S, Tang N, et al. WNT inhibitory factor 1 promoter hypermethylation is an early event during gallbladder cancer tumorigenesis that predicts poor survival. Gene 2017;30(622):42–9.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
ARTICLE IN PRESS Digestive and Liver Disease xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Digestive and Liver Disease journal homepage: www.elsevier.com/locate/dld
Meta-Analysis
Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis Edith Lahner ∗ , Marina Capasso, Marilia Carabotti, Bruno Annibale Medical-Surgical Department of Clinical Sciences and Translational Medicine, Sapienza University of Rome, Italy
a r t i c l e
i n f o
Article history: Received 11 January 2018 Received in revised form 16 May 2018 Accepted 17 May 2018 Available online xxx Keywords: Autoimmune gastritis Biliary tract cancer Cancer risk Leukaemia Lymphoma Meta-analysis Pernicious anaemia Systematic review
a b s t r a c t
Background: Pernicious anaemia (PA) is associated with increased gastric cancer risk, but the evidence is conflicting regarding the associated risk of other cancers. Aim: To systematically determine the incidence rates of gastro-intestinal cancers other than gastric cancers (GI-other-than-GC) and non-gastrointestinal cancers (non-GIC) in PA adults, globally and per tumour site, and the risk associated with PA for GI-other than GC and non-GIC. Methods: Studies of PA patients reporting the incidence of GI-other-than-GCs and non-GICs were identified with MEDLINE (PubMed)-EMBASE (from first date available to April 2017). A meta-analysis of annual cancer incidence rates was performed. The outcome was the cumulative incidence of GI-other-than-GCs and non-GICs (ratio between the numbers of new cancer cases identified during the follow-up period and the number of PA patients) and the incidence rate expressed as the rate of events-per-time-unit (person-years). Results: Of 82,257 PA patients, the pooled incidence rates/100 person-years for non-GCs and non-GICs of 0.27 (95% CI:0.16–0.42) and 0.23 (95% CI:0.22–0.25) were calculated by meta-analysis. Compared to the GLOBOCAN data for the general population from the countries of the included studies, the meta-analysis showed an overall relative risk (RR) of cancer in PA of 0.68 (95% CI:0.48–0.95). PA patients had a lower RR of colorectal, breast, liver, oesophageal, lung, thyroid, ovary, non-melanoma skin and kidney cancers but had a higher RR of biliary tract cancer (1.81:1.21–2.70), multiple myeloma (2.83:1.76–4.55), Hodgkin’s lymphoma (3.0:1.35–6.68), non-Hodgkin’s lymphoma (2.08: 1.58–2.75), and leukaemia (1.56:1.16–2.12). Conclusion: An overall lower RR of cancers-other-than-gastric-cancer in PA patients but an increased RR of biliary tract cancers and haematological malignancies was observed. © 2018 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
1. Introduction Pernicious anaemia (PA) is a macrocytic anaemia that results from cobalamin deficiency due to intrinsic factor deficiency, a protein avidly bound to dietary vitamin B12 that promotes its transport to the terminal ileum for absorption. Intrinsic factor deficiency is due to the presence of atrophic gastritis (AG), resulting in the loss of oxyntic mucosa and parietal cells and therefore the lack of chlorhydric acid as well as intrinsic factor secretion [1,2]. PA is considered a late stage of autoimmune gastritis, and it is often associated with positivity to parietal cell and/or intrinsic factor autoantibodies and
∗ Corresponding author at: Sapienza University of Rome, Department of MedicalSurgical Sciences and Translational Medicine, Via Grottarossa 1035, 00189 Rome, Italy. E-mail address: [email protected] (E. Lahner).
with the presence of other autoimmune disorders, usually autoimmune thyroid disease, type I diabetes, or vitiligo [3,4]. Similar to other autoimmune diseases, PA has been associated with an increased risk of cancer development [5]. A systematic review revealed a pooled gastric cancer incidence rate in PA patients of 0.3/100 person-years and an estimated 7-fold relative risk of gastric cancer [6]. In AG patients, type 1 gastric carcinoids may also occur [7–9]. High serum gastrin levels are found in PA as a response to damaged oxyntic mucosa and impaired gastric acid secretion. Hypergastrinaemia was shown to stimulate the growth of epithelial cells and to prevent apoptosis, possibly contributing to increased cancer risk [10]. In addition to gastric cancer and type 1 gastric carcinoids, PA has also been reported to be associated with an increased risk of cancers other than gastric neoplasias. Prior studies evaluating the possible association of PA with cancer risk have resulted in conflicting evidence; population-based studies mainly reported an increased odds ratio
https://doi.org/10.1016/j.dld.2018.05.012 1590-8658/© 2018 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
2
for haematological malignancies, such as monoclonal gammopathy of undetermined significance (MGUS) [11], multiple myeloma [12–14], acute myeloid leukaemia [14,15], chronic lymphocytic leukaemia [16], and myelodysplastic syndrome [14,15]. In contrast, a case-control study did not observe an association between PA and haematological malignancies [17]. Murphy et al. recently reported that patients with PA seem to be at increased risk for tonsillar cancer (OR 2.18), hypopharyngeal cancer (OR 1.92), oesophageal squamous cell carcinoma (OR 2.12), small intestinal cancer (OR 1.63) and liver cancer (OR 1.49) [14]. The aim of the present study was to determine the incidence rates of gastro-intestinal cancers other than gastric cancers (GIother than GC) and non-gastrointestinal cancers (non-GIC) in adult patients with PA, both globally and per tumour site, and to determine the risk associated with PA for GI-other than GC and non-GIC by systematically reviewing the literature. 2. Methods 2.1. Literature search strategy The literature on adult PA patients and the development of GI-other than GC and non-GIC was systematically reviewed. The search was conducted according to PRISMA guidelines [18]. The electronic databases MEDLINE (PubMed) and EMBASE were systematically searched combining the main keywords “pernicious anaemia”; “autoimmune gastritis”; and “cancer”. The search was translated into the following query: (“pernicious anaemia”[All Fields] OR “anaemia”, “pernicious”[MeSH Terms] OR (“anaemia”[All Fields] AND “pernicious”[All Fields]) OR (autoimmune[All Fields] AND (“gastritis”[MeSH Terms] OR “gastritis”[All Fields]) AND (“neoplasms”[All Fields] OR “cancer”[All Fields]) OR “carcinoma”[All Fields]) AND (“humans”[MeSH Terms] AND (English[lang] OR French[lang] OR German[lang] OR Spanish[lang]). No publication date restrictions were imposed. Reports published in English, German, French, Italian, and Spanish were considered. 2.2. Study selection Studies published from the first date available up to April 27th, 2017 were included in the systematic review if they fulfilled all of the following criteria: (1) an observational study including patients with PA and reporting the numbers of non-gastric and/or non-GI cancers identified during a defined follow-up period; (2) a study performed in adult patients; and (3) a study with an original full paper that presented unique data. Studies were excluded if (1) they were reviews, letters, editorials, or case-reports and (2) follow-up data were not available. Potentially relevant articles were screened for eligibility independently in an unblinded, standardized manner by the two reviewers (EL, MC), first by the abstract and then by the full text when necessary to determine whether they met the inclusion criteria. Disagreement between reviewers was resolved by discussion. The reference lists of the identified articles and relevant reviews were manually searched for additional studies that may have been missed using the electronic search strategy.
calculation of person-years (PY), age of patients (median or mean or range), gender, methods for identification of cancer, type of cancers, and numbers of cancer cases. 2.4. Outcome The outcome measures of interest were the cumulative incidence of GI-other than GCs and non-GICs, calculated as the ratio between the numbers of new cancer cases identified during the follow-up period and the number of PA patients, and the incidence rate was expressed as a rate of events per unit of time (personyears). 2.5. Statistical analyses From each included study, the number of incident non-GC and non-GIC cases and the number of PA patients exposed to risk were extracted. The cumulative incidence and the PY incidence rates in PA patients were calculated by the reviewers, if not explicitly stated. The cumulative incidence was calculated as the ratio between new non-GC and non-GIC cases and the number of PA patients. The PY incidence rates were calculated as the ratio between new non-GC and non-GIC cases and PY. The weighted summary proportion (pooled proportion) under the fixed and random effects model were calculated by a Freeman–Tukey transformation for all the non-GC and non-GIC cases and were calculated separately for specific cancer sites in single systems or organs. In the case of a positive “heterogeneity test” (p-value < 0.10), the more appropriate random effects model was taken into consideration, in which both the random variation within the studies and the variation between the different studies was incorporated [19]. The extent of heterogeneity was investigated using Cochran’s Q and I2 statistic [20]. The pooled incidence rates derived by meta-analysis were then compared with the annual cancer incidence rates of both genders in the general population aged over 40 years, as reported for the single European countries by the website GLOBOCAN (taking into consideration those countries to which the single studies included in the meta-analysis refer to) [21] by the Mantel–Haenszel method for calculating the weighted summary relative risk under the fixed effects model and the random effects model [19]. The statistical analysis was carried out using a dedicated software package (MedCalc Software, Mariakerke, Belgium, version 12.7.8.0). 2.6. Quality assessment The quality of all the included studies was evaluated based on the Newcastle-Ottawa quality assessment scale [22]. This scale awards a maximum of nine stars to each study. Three categories are considered: (i) cohort selection including four items, (ii) outcome assessment with three items, and (iii) comparability between cohorts and controls with two items. A study can be awarded a maximum of one star for each numbered item within the selection and outcome assessment categories and two stars for comparability. Studies were defined as of high quality when they obtained nine stars, of medium quality when they obtained seven or eight stars, of low quality when they obtained five or six stars, and of very low quality when they obtained four stars or less. Discrepancies in quality assessment were discussed and resolved by two reviewers (EL, MC).
2.3. Data extraction 3. Results Two reviewers (EL, MC) independently extracted the following information from each publication: name of the first author, publication year, country of study location, study design, criteria for diagnosis of PA, sources of participant selection, numbers of investigated patients, duration of follow-up period (expressed in years),
3.1. Search results The electronic search strategy identified a total of 2243 records from electronic databases and manual searches of the reference
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
3
Fig. 1. Flow chart of the study selection.
lists (Fig. 1). These articles were screened on the basis of the title and abstract. After application of the inclusion and exclusion criteria, 97 articles were retrieved for full-paper evaluation. Of these 97 fullpapers, 20 met the eligibility criteria and were subjected to data extraction [23–42]. 3.2. Quality assessment Details of the quality assessments of the included studies are given in Supplementary Table S1 in the online version at DOI: 10.1016/j.dld.2018.05.012. Of the twenty included studies, four (20.0%) [26,29,39,41] were of very low quality, five (25.0%) [23,24,28,30,33] of low quality, six (30.0%) [25,31,32,37,40,42] of medium quality and five (25.0%) [27,34–36,38] of high quality, according to the Newcastle-Ottawa quality assessment scale. 3.3. Characteristics of the included studies Female gender was prevalent in nine studies, with a range between 51.3% and 72.7%. In 1 study, there were fewer females than males (44.3%); in two studies, only male patients were investigated; and in the remaining studies, data on gender were lacking. The median age of patients was 71.4 years in all studies, ranging from 56.4 37–79.0 28 years.
The majority of the twenty studies were conducted in European countries, including nine (45.0%) in Northern Europe (Denmark, Finland, Sweden) and five (25.0%) in other European countries (United Kingdom, Germany). Five studies (25.0%) were American, and only one study (5.0%) was performed in China. Ten (50%) studies had a prospective design. The diagnostic criteria for PA diagnosis were thoroughly reported in ten (50.0%) studies. In nine studies, PA diagnosis was made by blood tests (macrocytic anaemia, low serum cobalamin or iron levels, hypergastrinaemia, low pepsinogen I levels), positive Schilling test, gastric pH, or megaloblastic bone marrow [23,24,26,28,33,37,39–41]; in seven studies, the data were extracted based on International Classification of Disease (ICD) [27,32,34–36,38,42]; and in one study, were based on hospital records [31]. In three studies [25,29,30], further details for the included PA patients were not provided. Most of the studies included patients who were first hospitalized for PA and afterwards were followed-up as outpatients. Three studies included outpatients referred to haematological units for anaemia or local healthcare providers [38–40]. In one study [25], patients were recorded in the files of eight PA clinics, but data were not provided on whether the patients were outpatients or hospitalized. Most of the studies retrieved the diagnosis of GI and non-GI cancers by biopsies using ICD codes from National Registers: three from local computer-based files [28,35,40] and four studies from
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
4
Table 1 Cancer incidence per organ system in pernicious anaemia (PA) patients. Cancer cases not included: 1 kidney/lymphoma/neurinoma [24]; 3 genital (not specified if female or male from 27); 24 not specified [26]; 26 not specified [27]; 44 not specified [25]; 1 metastasis [29]. Organ system
Cases of cancer/PA patients (cumulative incidence%) [annual incidence rate per 100.000 person-years]
Organ
Cases of cancer/PA patients (cumulative incidence%) [annual incidence rate per 100.000 person-years]
References
CNS
25/15,155 (0.2) [5.7]
Brain and meninges
25/15,155 (0.2) [5.7]
[24,27,28,34,36]
Maxillo-facial
206/30,042 (0.7) [8.4]
Sinus maxillary Eye Buccal cavity and pharynx
2/276 (0.7) [17.3] 1/1561 (0.06) [2.8] 203/30,004 (0.7) [9.1]
[24,29] [27] [27–29,32,34–36]
Respiratory
180/15,575 (1.2) [13.6]
Larynx Lung Other
2/4555 (0.04) [3.4] 165/15,537 (1.1) [13.6] 13/5161 (0.2) [37.0]
[24,34] [27–29,34–36,39] [27]
Endocrine
12/10,190 (0.1) [2.4]
Thyroid Other
5/5673 (0.1) [2.0] 7/4517 (0.1) [26.3]
[23,24,27,37,39] [34]
GI other than gastric
767/67,654 (1.1) [8.2]
Esophagus Small Bowel Colon Rectum Anus Liver Biliary tract Pancreas Other
141/50,902 (0.3) [3.6] 19/24,753 (0.08) [1.1] 271/30,789 (0.9) [10.3] 121/30,036 (0.4) [4.8] 3/11,839 (0.02) [4.4] 40/18,589 (0.2) [4.2] 30/15,478 (0.2) [6.7] 123/33,483 (0.4) [4.8] 19/5522 (0.4) [24.9]
[27,29,32,34–36,42] [27–29,32,35,36] [23,26,27,28,32–37,40] [24,26,27,29,32,34–36] [32] [26–28,34–37] [26–28,34,36] [26,27,29,34–36,38] [26,27]
Female reproductive
126/15,656 (0.8) [11.9]
Breast Uterus Ovary
79/15,656 (0.5) [7.5] 31/10,495 (0.3) [4.9] 16/9827 (0.2) [3.5]
[27–29,34,36,37] [28,29,34,36,37] [29,34,36]
Male reproductive
191/15,189 (1.3) [31.9]
Prostate
191/15,189 (1.3) [31.9]
[27,33,34,36,37]
Urinary
101/20,149 (0.5) [8.4]
Kidney Bladder
41/14,988 (0.3) [5.2] 60/14,988 (0.4) [7.6]
[26,29,34,36] [27,29,34,36]
Hematopoietic
199/29,713 (0.7) [10.0]
Hodgkin Lymphoma Non Hodgkin Lymphoma Multiple Myeloma Leukemia Other
3/9678 (0.03) [2.4] 70/22,240 (0.3) [15.4] 28/14,888 (0.2) [6.5] 76/16,557 (0.5) [10.8] 22/11,808 (0.2) [6.1]
[27,34] [31,34,36] [27,33–36] [25,27,30,34,36,39] [25,27,28,33,34]
Skin
86/14,750 (0.6) [32.8]
Melanoma Non melanoma
21/14,750 (0.1) [8.0] 65/9589 (0.7) [61.6]
[27,34,36] [34,36]
Bone and soft tissue
13/5766 (0.2) [4.9]
Bone Connective/soft tissue
6/5161 (0.1) [17.1] 7/5766 (0.1) [2.6]
[27] [27–29]
clinical interviews or death certificates [25,37,39,41]. Three studies did not specify how the cancer diagnoses were assessed [29,30,33]. For the data on cancer incidence, four studies reported only GI-other than GC cases [38,40–42], and three studies reported only non-GIC cases [30,31,39], while the remaining studies reported both cases. The characteristics of the included studies are summarized in Supplementary Tables S2 and S3 in the online version at DOI: 10.1016/j.dld.2018.05.012. 3.4. Cancer incidence With regard to the cancer incidence in PA patients, we found that the overall cumulative incidence was 2.4% for a cumulative follow-up of 185,5 years, ranging from 0.2% [42] to 15.8% [24], with 1993 new cancer cases in a total PA population of 82,257 patients (15,285,818.3 PY), corresponding to an overall incidence rate of 13.0/100,000 PY. In particular, the cumulative cancer incidence ranged from 0.2% [38,42] to 9.2% [34] in the 9 Northern European studies, from 0.2% [38] to 8.3% [41] in the 5 studies of other European countries, from 5.1% to 15.8% in the 5 American studies and 6.0% in the Chinese study. Overall, a total of 67,654 (9,392,404.8 PY) and 45,373 (6,825,006.7 PY) PA patients were studied, with 767 and 1194 new cancer cases in the GI-other than GC and non-GIC group, respectively. Examining GI-other than GCs, the cumulative incidence was 1.1% for a cumulative follow-up of 138.8 years (incidence rate
8.2/100,000-PY), ranging from 0.2% [38,42] to 8.0% [26] (Supplementary Table S2 in the online version at DOI: 10.1016/j.dld. 2018.05.012). Examining non-GICs, the cumulative incidence was 2.6% for a cumulative follow-up of 150.2 years (incidence rate 17.5/100,000-PY), ranging from 0.3% [32] to 13.2% [24] (Supplementary Table S3 in the online version at DOI: 10.1016/j.dld.2018. 05.012). The heterogeneity between studies was statistically significant in both groups (p < 0.0001). From the meta-analysis, a pooled incidence rate/100-PY for non-GCs and non-GICs of 0.27 (95% CI: 0.16–0.42) (Supplementary Fig. S1 in the online version at DOI: 10. 1016/j.dld.2018.05.012) and 0.23 (95% CI: 0.22–0.25) (Supplementary Fig. S2 in the online version at DOI: 10.1016/j.dld.2018.05.012) was found. Analysing the specific site of cancer incidence (Table 1), we found that the highest incidence rates/100,000-PY were observed for skin cancer (32.8, driven by non-melanoma cancer: 61.6) and prostate cancer (31.9). A meta-analysis performed using the annual cancer incidence rates per organ showed an overall relative risk (RR) for cancer in PA compared to the country-specific GLOBOCAN data of the general population (adults over 40 years of age) of 0.68 (95% CI: 0.48–0.95). With regard to site-specific cancers, PA patients had a lower RR (RR: 95% CI) compared to the general population for lung (0.26: 0.21–0.31), thyroid (0.34: 0.21–0.57), oesophagus (0.26: 0.18–0.37), colorectal (0.14: 0.01–0.19), liver (0.18: 0.13–0.24), pancreas (0.65: 0.45–0.93), breast (0.17: 0.13–0.22), ovary (0.45: 0.30–0.67), prostate (0.72: 0.63–0.84), kidney (0.61: 0.43–0.86),
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
5
Fig. 2. Relative risk of cancer per organ in patients with pernicious anaemia compared with the GLOBOCAN data for the general population (adult patients over 40 years of age) from the countries in which the included studies were performed. p < 0.0001; I2 = 96.79%; 95% CI: 95.92–97.47.
and non-melanoma skin cancers (0.56: 0.50–0.61). In contrast, PA patients had a higher risk than the general population for biliary tract cancer (1.81: 1.21–2.70) and for haematological malignancies, such as multiple myeloma (2.83: 1.76–4.55), Hodgkin’s lymphoma (3.0: 1.35–6.68), non-Hodgkin’s lymphoma (2.08: 1.58–2.75), and leukaemia (1.56: 1.16–2.12) (Fig. 2).
4. Discussion Evidence on overall cancer incidence in PA is sparse and conflicting. To our knowledge, this is the first systematic review of the estimate of GI-other than GCs and non-GICs cumulative incidence and incidence rates of PY in these patients. From 2243 articles retrieved from a search strategy with no time limit, we selected 20 papers according to our inclusion criteria published between 1950 and 2014, with a total of 82,257 PA patients (15,285,818.3 PY). Overall, 1993 new cases of cancer other than gastric cancer were identified, corresponding to a cumulative incidence of 2.4%, ranging from 0.2% to 15.8%. Furthermore, 767 and 1194 new cases of GI cancers other than GCs and non-GICs, respectively, were identified, with a cumulative incidence of 1.1% (incidence rate 8.2/100,000 PY) and 2.6% (15.7/100,000 PY), respectively. Compared with the GLOBOCAN data for the general population over 40 years of age (from the countries in which the studies were performed), PA patients had an overall lower RR (0.68) of developing cancers (other than gastric), even if we found an increased risk for haematological malignancies (in particular multiple myeloma [2.83] and Hodgkin’s lymphoma [3.0], and biliary tract cancers [1.8]). Subjects with PA have long been suggested to have an increased risk of gastric cancer as well as of gastric carcinoid tumours [6,8,9,30,43], but PA has been reported to be associated with an increased risk of other cancers. However, the paucity of previous studies on PA and cancer make this association controversial [10–15,17,35]. In our analysis, we found that PA patients had a lower RR of GI-other than GCs (oesophagus, liver, pancreas and colorec-
tum). The possible association between PA and GI-other than GCs has been debated in the literature, although the evidence is conflicting. An inverse relationship between gastric atrophy and oesophageal cancer has been observed in some studies [44–46], which is thought to reflect reduced acid reflux from the stomach to the oesophagus as a consequence of the oxyntic mucosa atrophy and the loss of parietal cells. In agreement with this finding, Murphy et al. found a positive association with oesophageal squamous cell carcinoma but no association with oesophageal adenocarcinoma [14]. As the authors stated, this relationship is not fully understood and may or may not be causal [14]. Gastrin and its precursors activate several pathways important in tumourigenesis, such as the beta catenin, MAP kinase and JAK2/STAT3 pathways, and they were described in association with colorectal, pancreatic and liver cancers in addition to gastric cancers and carcinoids [26,40,47]. Gastrin, at physiological concentrations, was shown to act as a growth factor for colorectal cancer through endocrine as well autocrine/paracrine mechanisms, and pre-malignant adenomas were also shown to express an isoform of the cholecystokinin B/gastrin receptor [48,49]. However, the lack of association between PA and colorectal cancer risk was recently demonstrated in a case control study [50] as well as in the largest population study currently available [10]. Only one study reported an increased risk of liver cancer for subjects with PA [34], which was recently confirmed in females [14]. No specific association of PA with lung, breast, prostate or kidney cancers was found in a recent population-based study [14], supporting our results. Thus, the findings of the present systematic review are in agreement with these data, denying the higher risk of cancers in these organs in patients with PA. However, in our analysis, we found an increased relative risk for haematological malignancies and biliary tract cancers. The role for immune dysregulation in haematological malignancies is well established, and several studies investigated a possible association between PA and haematological malignancies. Overall, PA seemed to increase the risk for multiple myeloma [12–14], acute myeloid
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7 6
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
leukaemia and myelodysplastic syndrome [14,15]. However, the association between PA and multiple myeloma was not confirmed by Lewis et al. [51] or Lindqvist et al. [11]. In addition, the study by Soderberg et al. [17] did not find any association between PA and all haematological diseases. The association between the risk of haematological malignancies and PA could be related to cobalamin deficiency [12,15]. Cobalamin deficiency may promote the development of cancer by causing derangement of one-carbon metabolism nutrients (folate, vitamin B6 , riboflavin, homocysteine) associated with aberrations in DNA methylation and DNA synthesis, as proposed by some authors [52,53]. When cobalamin is insufficient, cellular folate accumulates in the methylfolate form but cannot be metabolically utilized as a consequence of the block in the methionine synthase reaction, creating a functional folate deficiency referred to as the “methylfolate trap” [54]. Previous studies in bone marrow cells and immortalized cultured cells [55,56] demonstrated that cobalamin deficiency increased uracil misincorporation into DNA, suggesting that this functional folate deficiency by cobalamin deficiency results in inadequate thymidylate synthesis. Some authors observed that localized cobalamin deficiency in squamous-cell lung cancer tissue was associated with genomic DNA hypomethylation, suggesting that decreased availability of methylfolate by cobalamin deficiency subsequently reduces genomic DNA methylation [52]. We also found a positive association of PA with the risk of biliary tract cancers. There are no specific literature data on this association; however, a possible role for DNA hypomethylation can be hypothesized, considering that aberrant DNA methylation is an epigenetic mechanism that can occur early in carcinogenesis [57]. A recent molecular study demonstrated the up-regulation of Dicer in cholangiocarcinoma cells, promoting their proliferation and invasion [58]. Dicer is known to regulate methylation of CpG islands in mammalian cancer cells [59,60], interacting with heterochromatin protein 1␣ (HP1␣). DNA methylation is also considered one of the key molecular mechanisms in gallbladder tumourigenesis [61]. In a recent study, the aberrant methylation of the WIF-1 promoter was found to be involved in the malignant transformation of gallbladder cancer [62]. We are aware that this study has some limits. To evaluate the quality of included studies, we applied the Newcastle-Ottawa Quality Scale for cohort studies, which permits a critical appraisal of data [22]. We found that almost half of the included studies were of high/medium quality. In spite of this, clinical and methodological heterogeneity was present among the 20 studies included, resulting in a wide range of observed cumulative cancer incidence (0.2%–15.8%). The different approaches across the studies should take into account the decades in which the studies were performed. Diagnostic tools for both PA and cancer diagnosis have changed over time. A cancer diagnosis obtained from death certificates in the older studies could be not directly comparable with those obtained from ICD codes in National Cancer Registers. Additionally, six studies did not specifically report how the cancer was detected. Moreover, we excluded from this analysis the incidence rate for gastric cancer, as this issue was previously reported [6]. We know that the introduction of GI endoscopy has been a revolution in the diagnosis of upper GI diseases. This observation might impact the number of oesophageal cancers reported in older studies; some of those cases may include primary misdiagnosed gastric cancer. Notwithstanding these limits, we performed a meta-analysis based on the calculated GI-other than GCs and non-GICs incidence rates in PA patients. A further limitation of this systematic review is that for the comparison with the general population used for the calculations of the relative risk of cancer in PA, data from cancer registries (GLOBOCAN) were used, but the retrieved studies did not report control data, and this was a way to obtain estimates on the association of the single cancers with PA. To reduce this limitation, we used
as a comparison the GLOBOCAN data from adult patients over 40 years of age from those countries in which the studies included in the meta-analysis were performed. Although we are aware that the results of this meta-analysis may be biased by the limits explained above, in particular the structural fragility of the available information, we do think that they are acceptable as an estimate of the overall probability of PA patients to develop cancer and may represent a starting point to design further population-based prospective studies. In conclusion, this systematic review revealed an overall cancer risk (other than gastric cancer) for PA patients that was lower than the risk in the general population, except for haematological malignancies and biliary tract cancers. Further high-quality studies are needed to confirm the higher risk of these specific cancers in PA patients. Conflict of interest None declared. Funding This paper was funded in part by a grant from Sapienza University, funds of Oncology Doctorate 2016. Acknowledgments All the authors approved the final version of the article, including the authorship list. We are indebted to Mimma Ariano, Ales Casciaro, and Teresa Prioreschi of the University Library, Sapienza University, Sant’Andrea Hospital for their kind assistance during the retrieval of full text of papers. References [1] Annibale B, Lahner E, Delle Fave G. Diagnosis and management of pernicious anemia. Curr Gastroenterol Rep 2011;13:518–24. [2] Babior BM. Erythrocyte disorders: anemias related to disturbance of DNA synthesis (megaloblastic anemias). In: Williams JW, Beutler E, Erslev AJ, Lichtman MA, editors. Hematology. 4th ed. New York: McGraw-Hill; 1998. p. 453–81. [3] Toh BH, Gleeson PA, Whittingham S, van Driel IR. Autoimmune gastritis and pernicious anemia. In: Rose NR, Mackay IR, editors. The autoimmune diseases. 3rd ed. Academic Press; 1998. p. 459–76. [4] Lahner E, Centanni M, Agnello G, Gargano L, Vannella L, Iannoni C, et al. Occurrence and risk factors for autoimmune thyroid disease in patients with atrophic body gastritis. Am J Med 2008;121:136–41. [5] Lahner E, Annibale B. Pernicious anemia: new insights from a gastroenterological point of view. World J Gastroenterol 2009;15:5121–8. [6] Vannella L, Lahner E, Osborn J, Annibale B. Systematic review: gastric cancer incidence in pernicious anaemia. Aliment Pharmacol Ther 2013;37:375–82. [7] Vannella L, Sbrozzi-Vanni A, Lahner E, Bordi C, Pilozzi E, Corleto VD, et al. Development of type I gastric carcinoid in patients with chronic atrophic gastritis. Aliment Pharmacol Ther 2011;33:1361–9. [8] Kokkola A, Sjöblom SM, Haapiainen R, Sipponen P, Puolakkainen P, Järvinen H. The risk of gastric carcinoma and carcinoid tumours in patients with pernicious anaemia. A prospective follow-up study. Scand J Gastroenterol 1998;33:88–92. [9] Lahner E, Esposito G, Pilozzi E, Purchiaroni F, Corleto VD, Di Giulio E, et al. Occurrence of gastric cancer and carcinoids in atrophic gastritis during prospective long-term follow-up. Scand J Gastroenterol 2015;50:856–65. [10] Boursi B, Mamtani R, Haynes K, Yang YX. Pernicious anemia and colorectal cancer risk — a nested case-control study. Dig Liver Dis 2016;48:1386–90. [11] Lindqvist EK, Goldin LR, Landgren O, Blimark C, Mellqvist UH, Turesson I, et al. Personal and family history of immune-related conditions increase the risk of plasma cell disorders: a population-based study. Blood 2011;118:6284–91. [12] Anderson LA, Gadalla S, Morton LM, Landgren O, Pfeiffer R, Warren JL, et al. Population-based study of autoimmune conditions and the risk of specific lymphoid malignancies. Int J Cancer 2009;125:398–405. [13] Brown LM, Gridley G, Check D, Landgren O. Risk of multiple myeloma and monoclonal gammopathy of undetermined significance among white and black male United States veterans with prior autoimmune, infectious, inflammatory, allergic disorders. Blood 2008;111:3388–94. [14] Murphy G, Dawsey SM, Engels EA, Ricker W, Parsons R, Etemadi A, et al. Cancer risk after pernicious anemia in the US elderly population. Clin Gastroenterol Hepatol 2015;13, 2282-9.e1-4.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012
G Model YDLD-3758; No. of Pages 7
ARTICLE IN PRESS E. Lahner et al. / Digestive and Liver Disease xxx (2018) xxx–xxx
[15] Anderson LA, Pfeiffer R, Landgren O, Gadalla S, Berndt SI, Engels EA. Risk of myeloid malignancies in patients with autoimmune conditions. Br J Cancer 2009;100:822–8. [16] Landgren O, Engels EA, Caporaso NE, Gridley G, Mellemkjaer L, Hemminki K, et al. Patterns of autoimmunity and subsequent chronic lymphocytic leukemia in Nordic countries. Blood 2006;108:292–6. [17] Soderberg KC, Jonsson F, Winqvist O, Hagmar L, Feychting M. Autoimmune diseases, asthma and risk of haematological malignancies: a nationwide casecontrol study in Sweden. Eur J Cancer 2006;42:3028–33. [18] Moher D, Liberati A, Tetzlaff J, Altman DG, PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med 2009;6:e1000097. [19] DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88. [20] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60. [21] GLOBOCAN 2012: Estimated Cancer Incidence, Mortality and Prevalence Worldwide in 2012. International Agency for Research on Cancer, World Health Organization. Available at: http://globocan.iarc.fr/old/factsheet.asp. [Accessed 24 July 2017]. [22] Wells GA, Shea B, O’Connell D, Peterson J, Welch V, Losos M, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of non randomised studies in meta-analyses. Available at: http://www.ohri.ca/programs/clinical epidemiology/oxford.asp. [Accessed 24 July 2017]. [23] Armbrecht U, Stockbrügger RW, Rode J, Menon GG, Cotton PB. Development of gastric dysplasia in pernicious anaemia: a clinical and endoscopic follow up study of 80 patients. Gut 1990;31:1105–9. [24] Arvanitakis C, Holmes FF, Hearne 3rd E. A possible association of pernicious anemia with neoplasia. Oncology 1979;36:127–9. [25] Blackburn EK, Callender ST, Dacie JV, Doll R, Girdwood RH, Mollin DL, et al. Possible association between pernicious anaemia and leukaemia: a prospective study of 1625 patients with a note on the very high incidence of stomach cancer. Int J Cancer 1968;3:163–70. [26] Borch K, Kullman E, Hallhagen S, Ledin T, Ihse I. Increased incidence of pancreatic neoplasia in pernicious anemia. World J Surg 1988;12:866–70. [27] Brinton LA, Gridley G, Hrubec Z, Hoover R, Fraumeni JF. Cancer risk following pernicious anaemia. Br J Cancer 1989;59:810–3. [28] Chan JC, Liu HS, Kho BC, Lau TK, Li VL, Chan FH, et al. Longitudinal study of Chinese patients with pernicious anaemia. Postgrad Med J 2008;84:644–50. [29] Demmler K. Causes of death in patients treated for pernicious anemia. Med Klin 1966;61:575–7. [30] Elsborg L, Andersen D, Bastrup-Madsen P. Gastrocamera screening in pernicious anaemia. With special reference to the occurrence of gastric polyps and cancer. Scand J Gastroenterol 1973;8:5–8. [31] Fallah M, Liu X, Ji J, Försti A, Sundquist K, Hemminki K. Autoimmune diseases associated with non-Hodgkin lymphoma: a nation wide cohort study. Ann Oncol 2014;25:2025–30. [32] Hemminki K, Liu X, Ji J, Sundquist J, Sundquist K. Autoimmune disease and subsequent digestive tract cancer by histology. Ann Oncol 2012;23:927–33. [33] Hoffman NR. The relationship between pernicious anemia and cancer of the stomach. Geriatrics 1970;25:90–5. [34] Hsing AW, Hansson LE, McLaughlin JK, Nyren O, Blot WJ, Ekbom A, et al. Pernicious anemia and subsequent cancer. A population-based cohort study. Cancer 1993;71:745–50. [35] Landgren AM, Landgren O, Gridley G, Dores GM, LInet MS, Morton LM. Autoimmune disease and subsequent risk of developing alimentary tract cancers among 4.5 million US male veterans. Cancer 2011;117:1163–71. [36] Mellemkjaer L, Gridley G, Møller H, Hsing AW, Linet MS, Brinton LA, et al. Pernicious anaemia and cancer risk in Denmark. Br J Cancer 1996;73:998–1000. [37] Mosbech J, Videbaek A. Mortality from and risk of gastric carcinoma among patients with pernicious anaemia. Br Med J 1950;2:390–4. [38] Shah P, Rhim AD, Haynes K, Hwang WT, Yang YX. Diagnosis of pernicious anemia and the risk of pancreatic cancer. Pancreas 2014;43:422–6.
7
[39] Siurala M, Varis K, Wiljasalo M. Studies of patients with atrophic gastritis: a 10-15-year follow-up. Scand J Gastroenterol 1966;1:40–8. [40] Talley NJ, Chute CG, Larson DE, Epstein R, Lydick E, Melton J. Risk for colorectal adenocarcinoma in pernicious anemia. A population-based cohort study. Ann Intern Med 1989;111:738–42. [41] von Knorre G, Pechau KG. Late fate of patients with pernicious anemia. Z Gesamte Inn Med 1975;30:701–6. [42] Ye W, Nyren O. Risk of cancers of the oesophagus and stomach by histology or subsite in patients hospitalised for pernicious anaemia. Gut 2003;52:938–41. [43] Sjöblom SM, Sipponen P, Järvinen H. Gastroscopic follow up of pernicious anaemia patients. Gut 1993;34:28–32. [44] Cook MB, Dawsey SM, Diaw L, Blaser MJ, Perez-Perez GI, Abnet CC, et al. Serum pepsinogens and Helicobacter pylori in relation to the risk of esophageal squamous cell carcinoma in the alpha-tocopherol, beta-carotene cancer prevention study. Cancer Epidemiol Biomarkers Prev 2010;19:1966–75. [45] Islami F, Sheikhattari P, Ren JS, Kamangar F. Gastric atrophy and risk of oesophageal cancer and gastric cardia adenocarcinoma—a systematic review and meta-analysis. Ann Oncol 2011;22:754–60. [46] Kamangar F, Diaw L, Wei WQ, Abnet CC, Wang GQ, Roth MJ, et al. Serum pepsinogens and risk of esophageal squamous dysplasia. Int J Cancer 2009;124:456–60. [47] Karlson BM, Ekbom A, Wacholder S, McLaughllin JK, Hsing AW. Cancer of the upper gastrointestinal tract among patients with pernicious anemia: a casecohort study. Scand J Gastroenterol 2000;35:847–51. [48] Singh P, Dai B, Wu H, Owlia A. A role of autocrine and endocrine gastrin-like peptides in colonic carcinogenesis. Curr Opin Gastroenterol 2000;16:68–77. [49] Smith AM, Watson SA. Gastrin and gastrin receptor activation — an early event in the adenoma-carcinoma sequence. Gut 2000;47:820–4. [50] Lahner E, Sbrozzi-Vanni A, Vannella L, Corleto VD, Di Giulio E, Delle Fave G, et al. No higher risk for colorectal cancer in atrophic gastritis-related hypergastrinemia. Dig Liver Dis 2012;44:793–7. [51] Lewis DR, Pottern LM, Brown LM, Silverman DT, Hayes RB, Schoenberg JB, et al. Multiple myeloma among blacks and whites in the United States: the role of chronic antigenic stimulation. Cancer Causes Control 1994;5:529–39. [52] Friso S, Choi SW. The potential cocarcinogenic effect of vitamin B12 deficiency. Clin Chem Lab Med 2005;43:1158–63. [53] Miranti EH, Stolzenberg-Solomon R, Weinstein SJ. Low vitamin B12 increases risk of gastric cancer: a prospective study of one-carbon metabolism nutrients and risk of upper gastrointestinal tract cancer. Int J Cancer 2017;141:1120–9. [54] Herbert V, Das KC. Folic acid and vitamin B12. In: Shils ME, Olson JA, Shike M, editors. Modern nutrition in health and disease, vol. 1. Philadelphia: Lea & Febiger; 1994. p. 402–25. [55] Wickramasinghe SN, Fida S. Misincorporation of uracil into the DNA of folateand B12-deficient HL60 cells. Eur J Haematol 1993;50:127–32. [56] Wickramasinghe SN, Fida S. Bone marrow cells from vitamin B12- and folatedeficient patients misincorporate uracil into DNA. Blood 1994;83:1656–61. [57] Esteller M. Epigenetics in cancer. N Engl J Med 2008;358:1148–59. [58] Cheng W, Qi Y, Tian L, Wang B, Huang W, Chen Y. Dicer promotes tumorigenesis by translocating to nucleus to promote SFRP1 promoter methylation in cholangiocarcinoma cells. Cell Death Dis 2017;8:e2628. [59] Ting AH, Suzuki H, Cope L, Schuebel KE, Lee BH, Toyota M, et al. A requirement for DICER to maintain full promoter CpG island hypermethylation in human cancer cells. Cancer Res 2008;68:2570–5. [60] Chen Y, Luo J, Tian R, Sun H, Zou S. miR-373 negatively regulates methylCpG-binding domain protein 2 (MBD2) in hilar cholangiocarcinoma. Dig Dis Sci 2011;56:1693–701. [61] Tekcham DS, Tiwari PK. Epigenetic regulation in gallbladder cancer: promoter methylation profiling as emergent novel biomarkers. Asia Pac J Clin Oncol 2016;12:332–48. [62] Lin B, Hong H, Jiang X, Li C, Zhu S, Tang N, et al. WNT inhibitory factor 1 promoter hypermethylation is an early event during gallbladder cancer tumorigenesis that predicts poor survival. Gene 2017;30(622):42–9.
Please cite this article in press as: Lahner E, et al. Incidence of cancer (other than gastric cancer) in pernicious anaemia: A systematic review with meta-analysis. Dig Liver Dis (2018), https://doi.org/10.1016/j.dld.2018.05.012