Association of Helicobacter [PMIDY24143873]

C orrespondence: J. Liu, Department of Cardiology, Tianjin Union Medicine Center and Tianjin People ’ s H ospital, Tianjin 300121, China. E-mail:liujuan_2013@ (Received 26 J une 2013 ; accepted 9 S eptember 2013) Scandinavian Journal of Infectious Diseases, 2013; Early Online: 1–9ISSN 0036-5548 print/ISSN 1651-1980 online © 2013 Informa Healthcare DOI: 10.3109/00365548.2013.844351O RIGINAL ARTICLEAssociation of Helicobacter pylori infection with diabetes mellitus and diabetic nephropathy: A meta-analysis of 39 studies involving more than 20,000 participantsF ENG W ANG 1 , J UAN L IU 2 & Z ONGSHUN L V 3F rom the 1 D epartment of Internal Medicine, Tianjin Union Medicine Center and Tianjin People ’ s Hospital, 2 D epartment ofCardiology, Tianjin Union Medicine Center and Tianjin People ’ s Hospital, and 3 D epartment of Gastroenterology, GeneralHospital of Tianjin Medical University, Tianjin, ChinaA bstractB ackground: Helicobacter pylori infects more than half of the world ’ s population. The aim of this study was to quantify the association between H. pylori infection and the risk of diabetes mellitus and diabetic nephropathy, and to detect at which stage the infection might have higher pathogenicity in the disease-free status – d iabetes mellitus – d iabetic nephropathy process.M ethods: A literature search was performed to identify studies published between 1997 and 2012 for relative risk estimates. Fixed and random effects meta-analytical techniques were conducted for diabetes mellitus and diabetic nephropathy. R esults: Thirty-seven case – c ontrol studies and 2 cohort studies were included. H. pylori was associated with an increased risk of each type of diabetes mellitus (odds ratio (OR) 2.00, 95% confi dence interval (CI) 1.82 – 2.20, p for heterogene-ity ϭ 0.07). The infection was also associated with increased risks of type 1 and type 2 diabetes mellitus, separately (OR 1.99, 95% CI 1.52 – 2.60, p for heterogeneity ϭ 0.15, and OR 2.15, 95% CI 1.81 – 2.55, p for heterogeneity ϭ 0.24, respec-tively). In addition, we found a signifi cant association between H. pylori infection and diabetic nephropathy risk (OR 1.60, 95% CI 1.10 – 2.33, p for heterogeneity ϭ 0.44). C onclusions: Our meta-analyses suggest a relationship between H. pylori infection and the risk of diabetes mellitus and diabetic nephropathy. The bacterium may be able to play its pathogenic role in the whole disease process, and this action may be stronger in type 2 diabetic patients than in type 1 diabetic patients.K eywords: D iabetes mellitus , t ype 1 diabetes mellitus , t ype 2 diabetes mellitus , H elicobacter pylori , n ephropathyI ntroduction Helicobacter pylori infects more than half of the world ’ s population [1]. In many regions, up to 80% of children younger than 10 y old are infected with this pathogen [2,3]. Over the past decades, compel-ling data have emerged suggesting that H. pylori can cause chronic gastritis, peptic ulcer disease, and gas-tric malignancies [4].M any studies have tried to determine the role of H. pylori in causation of diabetes mellitus and dia-betic complications, with confl icting results. Thus, there is still insuffi cient evidence to draw conclusions and to determine at what stage, before or after the onset of diabetes mellitus, H. pylori may act most in the disease-free status – d iabetes mellitus – d iabetic nephropathy process.We conducted a meta-analysis to examine the relationship between H. pylori infection and the risks of diabetes mellitus and diabetic nephropathy, and to identify at which stage this pathogen might have higher pathogenicity. M aterials and methodsD ata search Studies published from 1997 to 2012 were selected on the basis of a structured literature search in Medline and Embase. Search parameters were “ D ia-betes Complications [MeSH T erms] ” , “ D iabetes Mellitus [MeSH T erms] ” , “ D iabetes Mellitus, Type 2 [MeSH T erms] ” , “ D iabetes Mellitus, Type 1 [MeSH T erms] ” , “ D iabetic Nephropathies [MeSH T erms] ” ,2 F. Wang et al.“D iabetes [Title/Abstract]”,“D iabetic [Title/ Abstract] ”,“H elicobacter pylori [MeSH T erms] ”,“H elicobacter pylori [T ext Word] ”,“H. pylori [T ext Word] ”,“C ampylobacter pylori [T ext Word] ”,“C. pylori [T ext Word] ”, and their combinations. The references from articles were checked and suitable studies were also used. The date of the most recent search was 26 December 2012.W e performed all selections in duplicate. The fi nal inclusion of studies was determined by consen-sus, and when this failed, a third author adjudicated. The following inclusion criteria were used: (1) stud-ies that compared the incidence of diabetes mellitus or diabetic nephropathy in H. pylori infected people and uninfected people; (2) studies that compared the incidence of H. pylori infection in diabetes mellitus patients and non-diabetes mellitus controls. Studies that did not meet these inclusion criteria were excluded. Non-English articles were excluded.T wo trained research personnel independently extracted the following data in accordance with a pre-specifi ed protocol: fi rst author, country, year of publication, study size, study type, disease type, methods of diabetes mellitus diagnosis, methods of H. pylori detection, and adjusted variables. All data were double-entered.D ata analysisO ur meta-analysis was performed in line with the recommendations of the Cochrane Collaboration and the Quality of Reporting of Meta-analyses guidelines [5,6] and was conducted using Review Manager (RevMan) version 4.2 (The Cochrane Col-laboration, Software Update, Oxford, UK). We com-piled expected data (4-fold table data) of each study into a data matrix in RevMan 4.2. Pooling was per-formed by using both the fi xed effects method and random effects method. If results were homogeneous ( p for heterogeneity Ͼ0.05), the fixed effects model was reported. If not, the random effects model was reported. The effect measures estimated were the odds ratio (OR) and relative risk (RR) for dichoto-mous data reported with 95% confi dence intervals (95% CI). An OR/RR was considered statistically sig-nifi cant if the 95% CI did not include the value 1.T hree strategies were employed to quantitatively assess heterogeneity. First, graphical exploration with funnel plots was used to evaluate publication bias [7,8]. Second, we performed a sensitivity analysis. Each study was sequentially removed from the anal-ysis to determine its contribution to the overall effect size. Third, all studies were scored using the Newcas-tle –O ttawa Scale, with some modifi cations to match the needs of this study [9]. This scoring system eval-uated studies based on patient selection, comparabil-ity of the groups, and assessment of outcome. When an article received a score of more than 6 in this scoring system, it was regarded as a ‘h igh quality article ’.R esultsE ligible studiesO f the 298 studies found with the search parameters described above, 39 studies [10 –48] were eligible. The major reasons why other studies did not meet the inclusion criteria were that they were reviews/ comments/news/case reports with no original data ( nϭ75), non-English articles ( nϭ40), studies of other topics ( nϭ124), or studies of this area, but that had no expected data (4-fold table data) focusing on the relationship between H. pylori and diabetes mel-litus (or diabetic nephropathy) ( nϭ20). Searching the listed references by hand revealed no further study. The study characteristics are presented in T able I. The study by Chen et a l. [14] reported 2 sets of data. They were obtained separately from the National Health and Nutrition Examination Survey III and the National Health and Nutrition Examina-tion Survey 1999 –2000. Both were included in our meta-analysis.H. pylori versus diabetes mellitusT hirty-fi ve case –c ontrol studies including the article by Chen et a l. reported on the relationship between H. pylori and diabetes mellitus (T able I). Here, ‘d ia-betes mellitus ’was defined as any type of diabetes mellitus. All were considered for inclusion in our meta-analysis. The random effects pooled OR was 1.59 (95% CI 1.33 –1.90, pϽ0.00001). Another meta-analysis including 23 case –c ontrol studies that received high scores (score 7) was conducted, and the fi xed effects pooled OR was 2.00 (95% CI 1.82 –2.20, pϭ0.07) (Figure 1). Only 2 cohort studies [10,11] were found and a meta-analysis was con-ducted. The fi xed effects pooled RR was 1.68 (95% CI 0.93 –3.04, pϭ0.13).S ubgroup analysisT he data were stratifi ed by socioeconomic level, geo-graphical region, race, methods of diabetes mellitus diagnosis, and methods of H. pylori detection, and several meta-analyses were conducted again. The results are shown in T able II. We found no association between the infection and the disease risk in studies from East Asia (OR 0.65, 95% CI 0.39 –1.07, pϭ0.43).H elicobacter pylori and diabetes mellitus 3 T able I. Characteristics of studies included in the meta-analysis.First author Y Country Study size Disease type aDiabetesdiagnosis bBacteriadetection cMatchedvariables dQualitys coreH. pylori versus diabetes mellitus: cohort studies[10] Longo-Mbenza2012South Africa128/77 e DM131, 2, 3, 4, 5, 6, 77[11] Jeon2012USA719/63DM731, 2, 3, 5, 7, 8, 97H. pylori versus diabetes mellitus: case –c ontrol studies[12] T alebi-T aher2012Iran50/30 f DM2,31127[13] Oluyemi2012Nigeria100/100DM T2121, 2, 3, 97[14] Chen (1999) g 2012USA385/5687DM46127[14] Chen (III) g 2012USA585/6832DM43127[15] El-Eshmawy2011Egypt162/80DM T11, 231, 2, 10, 117[16] Agrawal2010India80/80DM T21, 241, 27[17] Sfarti2010Romania69/40DM T116127[18] Ibrahim2010Egypt98/102DM T2361, 26[19] Devrajani2010Pakistan74/74DM T212127[20] Krause2009Israel57/140DM T113126[21] Cabral2009Brazil15/30DM T116126[22] Ariizumi2008Japan67/67DM161, 26[23] Hamed2008Egypt80/60DM131, 2, 5, 77[24] Demir2008Turkey141/142DM T21, 261, 26[25] Bener2007Qatar210/210DM T21, 231, 27[26] Jaber2006Saudi Arabia61/543DM T113127[27] Gulcelik2005Turkey78/71DM T2111, 2, 5, 117[28] Gillum2004USA193/1628DM T213127[29] Altannavch2003Czech Rep52/47DM T2131, 2, 3, 87[30] Candelli2003Italy121/147DM T1151, 26[31] Maule2002Italy31/31DM T265127[32] Anastasios2002Greece67/105DM2, 311, 26[33] Cenerelli2002Italy30/43DM T2251, 2, 5, 87[34] Ko2001China63/55DM T2141, 26[35] Xia2001Australia429/170DM T1 T21326[36] Marrollo2001Italy74/117DM161, 27[37] Quatrini2001Italy71/71DM151, 27[38] Dore2000Italy385/506DM T1 T213106[39] Rosenstock2000Denmark58/2856DM53127[40] Arslan2000Turkey88/42DM T1631, 27[41] G üv ener1999Turkey51/25DM T26127[42] Salardi1999Italy103/236DM T16627[43] Gentile1998Italy164/164DM T2161, 2, 57[44] Gasbarrini1998Italy116/50DM T1151, 27[45] de Luis1998Spain80/100DM T1131, 2, 106[46] Pocecco1997Italy69/310DM T1139, 106H. pylori versus diabetic nephropathy: case –c ontrol studies[16] Agrawal2010India50/30 h DM T21, 241, 2, 5, 7, 117 [47] Ohnishi2008Japan70/60DM T2131, 3, 5, 6, 87[23] Hamed2008Egypt68/12DM13127[24] Demir2008Turkey87/54DM T21, 261, 26 [27] Gulcelik2005Turkey59/19DM T2112, 57[48] de Luis1998Spain53/74DM131, 2, 3, 5, 7, 106a D M, all types of diabetes mellitus; T1, type 1 diabetes mellitus; T2, type 2 diabetes mellitus.b1ϭm edical records; 2 ϭb ased on the American Diabetes Association criteria; 3 ϭr esearchers tested fasting plasma glucose; 4 ϭr esearchers tested HbA1c; 5 ϭs elf-reported; 6 ϭn ot mentioned; 7 ϭu sed several methods.c1ϭh istological staining; 2 ϭd etected antigens in stool samples; 3 ϭd etected antibodies by ELISA; 4 ϭr apid urease test; 5 ϭ13C-urea breath test; 6 ϭu sed several methods or used other rare method.d1ϭg ender; 2 ϭa ge; 3 ϭs moking; 4 ϭa lcohol intake; 5 ϭb ody mass index; 6 ϭh eart rate; 7 ϭb lood pressure; 8 ϭt otal cholesterol;9 ϭe ducation level; 10 ϭg eographic area; 11 ϭg astrointestinal symptoms; 12 ϭn ot mentioned.e N umber of participants with H. pylori infection/number of participants without H. pylori infection.f N umber of participants with diabetes/number of participants without diabetes.g C hen (III) ϭC hen, the National Health and Nutrition Examination Survey III; Chen (1999) ϭC hen, the National Health and Nutrition Examination Survey 1999 –2000.h N umber of diabetic patients with H. pylori infection/number of diabetic patients without H. pylori infection.4 F. Wang et al.F igure 1.Fixed effects meta-analysis of studies evaluating H elicobacter pylori infection and diabetes risk. The squares indicate point estimates of pathogenic effect, with the size of the square representing the weight attributed to each study, and 95% confi dence intervals (CI) are indicated by horizontal bars. The diamond represents the summary odds ratio (OR) from the pooled studies with 95% CI. ‘ C ases ’ represents diabetic patients; ‘ C ontrols ’ represents non-diabetic participants; n / N represents the number of diabetic patients with H. pylori infection/number of total diabetic patients; m / M represents the number of non-diabetic participants with H. pylori infection/number of total non-diabetic participants; ‘ T otal ’ represents the number of participants in each group.T able II. Results of the meta-analyses in the subgroups. SubgroupNo. of studies (study size a )OR (95% CI)b p -Valuec Socioeconomic levelDeveloped country 18 (3091/19,175)1.47 (1.18 – 1.83) Ͻ 0.00001Developing country 17 (1466/1816) 1.77 (1.27 – 2.48) Ͻ 0.00001Geographical region/race European 15 (1490/4823) 1.65 (1.23 – 2.20) Ͻ 0.0001Middle East 10 (1019/1305) 2.06 (1.47 – 2.88)0.004South Asia 2 (154/154) 2.53 (1.58 – 4.03)0.96East Asia2 (130/122)0.65 (0.39 – 1.07)0.43Immigration country3 (1592/14,317) 1.65 (1.21 – 2.27)0.0003Diabetes diagnosis Medical record25 (3011/5002) 1.52 (1.19 – 1.94) Ͻ 0.00001Based on ADA criteria 7 (740/690) 1.85 (1.26 – 2.72)0.01Bacteria detectionHistological staining4 (246/231) 2.59 (1.28 – 5.25)0.03Detected antibodies by ELISA 14 (2509/13,524) 1.57 (1.18 – 2.09) Ͻ 0.00001 13 C -urea breath test5 (369/342)1.37 (1.00 – 1.87)0.13O R, odds ratio; CI, confi dence interval; ADA, American Diabetes Association a N umber of participants with diabetes/number of participants without diabetes. b I f p for heterogeneity is Ͼ 0.05, the fi xed effects model is reported; if not, the random effects model is reported.c p -Value is the p -value for heterogeneity.H . pylori versus type 1/type 2 diabetes mellitus T hirteen case – c ontrol studies evaluated H . pylori infection and type 1 diabetes mellitus risk (T able I). The random effects pooled OR was 1.14 (95% CI0.70 – 1.85, p Ͻ 0.00001). After excluding 7 studies that received low scores (score 6), another meta-analysis was conducted and the fi xed effects pooled OR was 1.99 (95% CI 1.52 – 2.60, p ϭ 0.15) (Figure 2).H elicobacter pylori and diabetes mellitus5F igure 2. Fixed effects meta-analysis of studies evaluating Helicobacter pylori infection and 2 types of diabetes risk. The squares indicate point estimates of pathogenic effect, with the size of the square representing the weight attributed to each study, and 95% confi dence intervals (CI) are indicated by horizontal bars. The diamond represents the summary odds ratio (OR) from the pooled studies with 95% CI. ‘C ases ’represents diabetic patients; ‘C ontrols ’represents non-diabetic participants; n/N represents the number of diabetic patients with H. pylori infection/number of total diabetic patients; m/M represents the number of non-diabetic participants with H. pylori infection/ number of total non-diabetic participants; ‘T otal ’represents the number of participants in each group.S ixteen case –c ontrol studies evaluated the bacte-rial infection and type 2 diabetes mellitus risk (T able I). The random effects pooled OR was 1.78 (95% CI 1.42 –2.23, pϭ0.0006). One other meta-analysis only included the high scores (score 7) studies and the fi xed effects pooled OR was 2.15 (95% CI 1.81 –2.55, pϭ0.24) (Figure 2).H. pylori versus diabetic nephropathyS ix studies evaluated diabetic nephropathy risk in rela-tion to H. pylori infection (T able I). The fi xed effects pooled OR was 1.60 (95% CI 1.10 –2.33, pϭ0.44) (Figure 3). Four studies [16,24,27,47] focused on this relationship in the type 2 diabetic patients and a meta-analysis was conducted. The fi xed effects pooled OR was 1.88 (95% CI 1.19 –2.99, pϭ0.33). Two other studies [23,48] focused on the relationship in all types of diabetic patients, and the fi xed effects pooled OR was 1.15 (95% CI 0.60 –2.20, pϭ0.88).D iscussionT o our knowledge, the association of H. pylori infec-tion with diabetes mellitus is still controversial. Based on the results of our meta-analysis of case –c ontrol studies, we estimate an approximately 60 –100% dia-betes mellitus risk increase for the bacterial infection.H owever, type 1 and type 2 diabetes mellitus are quite different in pathogenesis. Our meta-analyses are consistent with this consensus. Based on the pres-ent results from the H. pylori versus type 1/type 2 diabetes mellitus group, a more significant type 2 diabetes mellitus risk increase for the bacterial infec-tion was found (approximately 80 –120%), but the relationship between type 1 diabetes mellitus and the bacterium is still not clear because the 95% CI in the meta-analysis of 13 case –c ontrol studies included the value 1.O ur meta-analyses suggest that H. pylori infec-tion in diabetic patients might lead to an approxi-mately 60% diabetic nephropathy risk increase, and this rate increased to nearly 90% in type 2 diabetic patients. However, no strong evidence was found to demonstrate such a relationship in type 1 diabetic patients. Therefore, the strength of the associations was similar in the ‘d isease-free status –d iabetes mel-litus ’stage and the ‘d iabetes mellitus –d iabetic neph-ropathy ’stage, suggesting that there might not be only 1 target point the bacteria focus on, and H.pylori is able to play its pathogenic role in the whole6 F. Wang et al.F igure 3. Fixed effects meta-analysis of studies evaluating Helicobacter pylori infection and diabetic nephropathy risk. The squares indicatepoint estimates of pathogenic effect, with the size of the square representing the weight attributed to each study, and 95% confi dence intervals (CI) are indicated by horizontal bars. The diamond represents the summary odds ratio (OR) from the pooled studies with 95% CI. n / N represents the number of diabetic nephropathy patients with H. pylori infection/number of diabetic patients with H. pylori infection; m / M represents the number of diabetic nephropathy patients without H. pylori infection/number of diabetic patients without H. pylori infection; ‘ T otal ’ represents the number of participants in each group.F igure 4. Funnel plots of studies evaluating Helicobacter pylori infection and diabetes and diabetic nephropathy risk.disease process, specifi cally in type 2 diabetic patients.T here are several reasons why our results are reli-able. First, all studies in our meta-analysis achieved good scores ( Ն 6) during quality assessment. More importantly, all results obtained from high quality articles (score 7) were calculated by the fi xed effects model. These studies showed better homogeneity and gave us more precise results. In our funnel plots, the largest studies were near the average, and small studies were spread on both sides of the average (Figure 4). A symmetric inverted funnel shape always arises from a ‘ w ell-behaved ’ dataset. Therefore, our graphical exploration with funnel plots showed no evidence of signifi cant publication bias in ourmeta-analysis.H elicobacter pylori and diabetes mellitus 7It should be remembered that H. pylori serology maynot equate with ongoing infection. Therefore, using this kind of diagnostic method has the potential to expand the relationship between the disease risk and the bacterial infection. Because an uncomfortable test process may result in the participants avoiding the use of histological methods, we could not rule out the possibility that the relationships in the stud-ies that used these methods had been weakened. Nevertheless our meta-analyses did not show that these 2 test methods (serological method and histo-logical staining method) led to any different results. Therefore it is unlikely that these limitations affectedour conclusions. In addition, the13 C -urea breath test is considered to be an accurate, comfortable test method, and the studies using this method gave us a positive result.C agA is considered to be a marker of the pres-ence of the cag pathogenicity island [51], which is associated with the severity of H. pylori-related dis-ease [52,53]. However, there were only a few studies focusing on the relationship between cagA-positive strain infections and disease risk. Therefore, a meta-analysis could not be conducted. Due to the lack of data, the association of H. pylori infection with other diabetic complications, such as neuropathy and isch-emic heart disease, was also not included in the meta-analysis. Many diabetic complications have a similar pathogenesis, so we hope our conclusions will be helpful for future research.T he mechanism by which H . pylori infection increases the risk of diabetes mellitus and its compli-cations remains to be elucidated. Several hypotheti-cal explanations have been suggested. Infl ammatory cytokine release may induce phosphorylation of ser-ine residues on the insulin receptor substrate, which may inhibit the interaction between the substrate and the insulin receptors, resulting in insulin dysfunction [54]. Lipopolysaccharides from the bacteria may lead to the activation of T oll-like receptors and con-sequent insulin resistance [55]. All these effects translate into decreased glycemic control and the occurrence of diabetes mellitus. Moreover, the presence of the bacterial infection also leads to microvascular insuffi ciency and the occurrence of atherosclerosis. Several cytokines, such as interleukin 6 and tumor necrosis factor alpha, may induce a direct endothelialitis, increase lipid synthesis, and trigger thrombosis [56]. Infl ammation caused by H. pylori may elevate production of lipoprotein(a), which has a vital role in atherothrombogenesis [57]. The direct invasion of the arterial wall by the bacte-ria may result in maturation of monocytes, and the latter produce proinfl ammatory cytokines. H. pylori infection may also stimulate the production of tissue-like factor, which promotes the conversion ofSecond, the only meta-analysis of cohort studies in our article reported that H. pylori infection did not increase the risk of diabetes mellitus [10,11]. This result was inconsistent with those of other case – c ontrol studies. Although ‘ r everse causality ’ and ‘ r ecall bias ’ might be potential limitations in the meta-analysis of case – c ontrol studies, we had enough reasons for choosing these cross-sectional studies in the meta-analysis. In the type 2 diabetes mellitus group, the mean age of the participants was higher than 45 y. Many studies suggest that most H. pylori infections occur in childhood and at a young age [49,50], and much earlier than the occurrence of diabetes mellitus. In all studies, the bacteria were detected by the researchers themselves using several reliable methods, and most studies established the diagnosis of diabetes mellitus from the medical records. Therefore, we do not believe that ‘ r everse causality ’ and ‘ r ecall bias ’ affected our conclusions. Furthermore, cohort studies focusing on this topic are rare, and only 2 studies involving less than 1000 participants were found. The small sample limited test performance. So, more prospective studies, espe-cially focusing on type 1 diabetes mellitus, should be conducted.T hird, several potential confounders such as age, gender, race, geographical region, socioeconomic level, and method of diabetes mellitus diagnosis could not be ignored. However, only 5 included stud-ies [14,28,35,36,39] gave us adjusted data (OR and 95% CI), and they separately focused on type 1 dia-betes mellitus, type 2 diabetes mellitus, and all types of diabetes mellitus. Therefore, a meta-analysis including adjusted data could not be conducted. However, we believe that this would not have affected our conclusions. Patients in most included studies were evenly matched for age and gender. In the sub-group analysis of the studies from developed coun-tries and developing countries separately, the same results were obtained (T able II). Two studies [22,34] from East Asia reported that the bacterial infection did not increase the risk of diabetes mellitus, which is in contrast to the studies from all other parts of the world; however the results of these 2 studies were limited by a small sample and lacked suffi cient reli-ability. Due to the lack of data, a better subgroup analysis to assess the effect of different disease diag-nosis methods on the validity of the fi ndings could not be conducted. However, 25 studies using medi-cal records and 7 studies using the American Diabe-tes Association criteria still gave the expected results.F ourth, in the meta-analysis, the included studies used several bacterial diagnostic tests. In many stud-ies, serum samples were analyzed for H. pylori IgGantibodies by enzyme-linked immunosorbent assay.8 F. Wang et al.fi brinogen to fi brin [58]. These effects may be related to the occurrence of several diabetic complications.I n conclusion, our meta-analyses suggest a rela-tionship between H. pylori infection and the risk of diabetes mellitus and diabetic nephropathy. The bac-terium may be able to play its pathogenic role in the whole disease process. This action may be stronger in type 2 diabetic patients than in type 1 diabetic patients.D eclaration of interest: N o confl ict of interest to declare.R eferencesB ener A,U duman S A ,A meen A,A lwash R,P asha M A ,[1]U smani M A ,e t a l .P revalence of Helicobacter pylori infec-tion among low socio-economic workers .J Commun Dis 2002 ;34 :179 –84 .D as J C ,P aul N.E pidemiology and pathophysiology of Heli-[2]cobacter pylori infection in children .I ndian J Pediatr 2007 ;74 :287 –90 .K halifa M M ,S haraf R R ,A ziz R K .H elicobacter pylori: a [3]poor man ’s gut pathogen? G ut Pathogens 2010 ;2:2.I ARC Working Group on the Evaluation of the Carcinogenic[4]Risks to Humans .H elicobacter pylori. IARC Monogr Eval Carcinog Risks Hum 1994 ;61 :177 –240 .C larke M,H orton R.B ringing it all together: Lancet –[5]Cochrane collaborate on systematic reviews .L ancet2001 ;357 :1728 .S troup D F ,B erlin J A ,M orton S C ,O lkin I,W illiamson G D , [6]R ennie D,e t a l .M eta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group .J AMA 2000 ;283 :2008 –12 .E gger M,D avey Smith G,S chneider M,M inder C.B ias in[7]meta-analysis detected by a simple, graphical test .B MJ1997 ;315 :629 –34 .E gger M,S mith G D .M isleading meta-analysis .B MJ[8]1995 ;311 :753 –4.A thanasiou T,A l-Ruzzeh S,K umar P,C rossman M C ,A mrani[9]M,P epper J R ,e t a l .O ff-pump myocardial revascularization is associated with less incidence of stroke in elderly patients .A nn Thorac Surg 2004 ;77 :745 –53 .L ongo-Mbenza B,N senga J N ,M okondjimobe E,G ombet T,[10]A ssori I N ,I bara J R ,e t a l .H elicobacter pylori infection isidentifi ed as a cardiovascular risk factor in Central Africans .V asc Health Risk Manag 2012 ;6:455 –61 .J eon C Y ,H aan M N ,C heng C,C layton E R ,M ayeda E R , [11]M iller J W ,e t a l .H elicobacter pylori infection is associated with an increased rate of diabetes .D iabetes Care 2012 ;35 :520 –5.T alebi-T aher M,M ashayekhi M,H ashemi M H ,B ahrani V.[12]H elicobacter pylori in diabetic and non-diabetic patientswith dyspepsia .A cta Med Iran 2012 ;50 :315 –8.O luyemi A,A nomneze E,S mith S,F asanmade O.P revalence [13]of a marker of active H elicobacter pylori infection among patients with type 2 diabetes mellitus in Lagos, Nigeria .B MC Res Notes 2012 ;5:284 .C hen Y,B laser M J .A ssociation between gastric Helicobacter[14]pylori colonization and glycated hemoglobin levels .J Infect Dis 2012 ;205 :1195 –202 .E l-Eshmawy M M ,E l-Hawary A K ,A bdel Gawad S S ,[15]El-Baiomy A A .H elicobacter pylori infection might be responsible for the interconnection between type 1 diabetesand autoimmune thyroiditis .D iabetol Metab Syndr 2011 ;3:28 .A grawal R P ,S harma R,G arg D,P okharna R,K ochar D K ,[16]K othari R P.R ole of Helicobacter pylori in causation of dia-betic gastropathies and non-gastrointestinal complications intype 2 diabetes .J Indian Med Assoc 2010 ;108 :140 –3.S farti C,T rifan A,H utanasu C,C ojocariu C,S ingeap A M , [17]S tanciu C.P revalence of gastroparesis in type 1 diabetes mel-litus and its relationship to dyspeptic symptoms .J Gastroin-testin Liver Dis 2010 ;19 :279 –84 .I brahim A,Z aher T,G honemy T A ,E l-Azim S A ,[18]El-Azim M A ,R amadan A.I mpact of cytotoxin-associated gene A of H elicobacter pylori strains on microalbuminuria in type 2 diabetes .S audi J Kidney Dis Transpl 2010 ;21 :694 –700 .D evrajani B R ,S hah S Z ,S oomro A A ,D evrajani T.T ype 2[19]diabetes mellitus: a risk factor for Helicobacter pylori infec-tion: a hospital based case –c ontrol study .I nt J Diabetes DevCtries 2010 ;30 :22 –6.K rause I,A naya J M ,F raser A,B arzilai O,R am M,A bad V,[20]e t a l .A nti-infectious antibodies and autoimmune-associatedautoantibodies in patients with type I diabetes mellitus and their close family members .A nn N Y Acad Sci 2009 ;1173 :633 –9.C abral V L ,P atr íc io F R ,G abbay M A ,D ib S A ,M iszputen S J .[21]I ntraepithelial lymphocytes in duodenum from Brazilianadolescents with type 1 diabetes .I nfl uence of Helicobacter pylori. Pediatr Diabetes 2009 ;10 :316 –20 .A riizumi K,K oike T,O hara S,I nomata Y,A be Y,I ijima K,[22]e t a l .I ncidence of refl ux esophagitis and H pylori infectionin diabetic patients .W orld J Gastroenterol 2008 ;14 :3212 –7.H amed S A ,A mine N F ,G alal G M ,H elal S R ,T ag El-Din[23]L M ,S hawky O A ,e t a l .V ascular risks and complications in diabetes mellitus: the role of helicobacter pylori infection .J Stroke Cerebrovasc Dis 2008 ;17 :86 –94 .D emir M,G okturk H S ,O zturk N A ,K ulaksizoglu M,[24]Serin E,Y ilmaz U.H elicobacter pylori prevalence in diabetesmellitus patients with dyspeptic symptoms and its relation-ship to glycemic control and late complications .D ig Dis Sci2008 ;53 :2646 –9.B ener A,M icallef R,Afi fiM,D erbala M,A l-Mulla H M ,[25]U smani M A .A ssociation between type 2 diabetes mellitus and H elicobacter pylori infection .T urk J Gastroenterol 2007 ;18 :225 –9.J aber S M .H elicobacter pylori seropositivity in children with [26]chronic disease in Jeddah, Saudi Arabia .S audi J Gastroen-terol 2006 ;12 :21 –6.G ulcelik N E ,K aya E,D emirbas B,C ulha C,K oc G,[27]Ozkaya M,e t a l .H elicobacter pylori prevalence in diabetic patients and its relationship with dyspepsia and autonomic neuropathy .J Endocrinol Invest 2005 ;28 :214 –7.G illum R F .I nfection with H elicobacter pylori, coronary[28]heart disease, cardiovascular risk factors, and systemic inflammation: the Third National H ealth and Nutrition Examination Survey .J Natl Med Assoc 2004 ;96 :1470 –6.A ltannavch Ts, R oubalov áK,B roz J,H rub áD,A nd e˘l M.[29]S erological markers of Chlamydia pneumoniae, cytomegalo-virus and Helicobacter pylori infection in diabetic and non-diabetic patients with unstable angina pectoris .C ent Eur J Public Health 2003 ;11 :102 –6.C andelli M,R igante D,M arietti G,N ista E C ,C rea F,[30]Bartolozzi F,e t a l .H elicobacter pylori, gastrointestinal symp-toms, and metabolic control in young type 1 diabetes mellitus patients .P ediatrics 2003 ;111 :800 –3.。