- Email: [email protected]
Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review
HLC 2521 1–8
REVIEW
Heart, Lung and Circulation (2017) xx, 1–8 1443-9506/04/$36.00 https://doi.org/10.1016/j.hlc.2017.10.020
Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review
1 2 3 4
[TD$FIRSNAME]Timothy P[2_TD$IF].[TD$FIRSNAME.] [TD$SURNAME]Shiraev[TD$SURNAME.] *[19_TD$IF], [TD$FIRSNAME]Andrew[TD$FIRSNAME.] [TD$SURNAME]Bullen[TD$SURNAME.]
5
Q1
6 7 8
Q2 Department of Vascular Surgery, St George Hospital, Kogarah, NSW, Australia Received 4 April 2017; received in revised form 15 October 2017; accepted 26 October 2017; online published-ahead-of-print xxx
Q4
Background
Proton pump inhibitors (PPIs) are a commonly prescribed medication that recent data has linked to an increased risk of cardiovascular morbidity and all cause morbidity. The current study sought to perform a systematic review to investigate the link between PPIs and morbidity and mortality
Methods
A systematic review was carried out as per the PRISMA guidelines, with information databases including Pubmed, Medline, and the Cochrane Review Database. English-language studies of all types published from January 1990 to October 2016 were considered. Dichotomous analysis generating odds ratios was performed using RevMan Version 5.3.
Results
Thirty-seven studies were considered, of which five directly compared the effect of PPI use on mortality and/or cardiovascular morbidity (including 22,427 patients in mortality datasets, and 354,446 patients in morbidity datasets). For patients taking PPIs, all cause mortality (OR 1.68 [95% CI 1.53–1.84), p [24_TD$IF]< 0.001) and rate of major cardiovascular events (OR 1.54 [95% CI 1.11–2.13], p = 0.01) were significantly higher.
Conclusions
The current systematic review demonstrates that, in patients using PPIs, there was a significant increase in morbidity due to cardiovascular disease. Careful consideration should be given to the prescription of PPIs while clinical equipoise remains. Further research in the area is required.
Q5
Q6
Keywords
Proton pump inhibitor Myocardial infarct Cardiovascular disease AMDA Nitric oxide Mortality
9 10 11
Introduction Q7
12 13 14 15 16 17 18 19 20 21
Q8
Proton pump inhibitors (PPIs) are indicated in gastroesophageal reflux disease (GORD), and are often the first line in anti-acid therapy as they have been demonstrated to be superior to H2 receptor antagonists [1]. Internationally, over 113 million PPI prescriptions are filled annually, accounting for $13 billion in sales [2]. However, PPIs are not free of side effects, with evidence for increased risks of diarrhoea, pneumonia and fractures in long-term users [1]. PPIs are often also prescribed to counteract the gastrointestinal complications of the antiplatelet aspirin, itself given to prevent atherosclerotic disease.
In the past decade, a number of studies have suggested there was a marked increase in cardiovascular morbidity and mortality in patients taking PPIs concomitantly with clopidogrel [25_TD$IF][3–5], to the extent that the FDA issued a formal warning about this interaction [6]. However, this has recently been suggested not to be a true interaction [7], especially as an increase in mortality and cardiac events has been seen even in patients not taking clopidogrel [8,9], and mechanisms for this increased thrombotic risk have recently been proposed [10]. Subsequent meta-analyses agree that there is an increase in adverse cardiac events in patients even not taking clopidogrel [26_TD$IF][11–16] and work by Sherwood et al. suggested that degree of cardiovascular risk
Q3 *Corresponding[23_TD$IF] author. Tel.: +61 9113 1111, Fax: +6191133984. Email: [email protected] Crown Copyright © 2017 Published by Elsevier B.V. on behalf of Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). All rights reserved.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
22 23 24 25 26 27 28 29 30 31 32 33
HLC 2521 1–8
2
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
T.P. Shiraev, A. Bullen
even differs between individual PPIs [12]. Debate remains as to whether PPIs do truly carry an increase in cardiovascular morbidity and mortality [27_TD$IF][17–19], and to what extent. We sought to investigate the impact of proton pump inhibitor use on both mortality, and morbidity due to thrombotic complications, namely ischaemic heart disease and vascular disease (including cerebrovascular accidents and limb ischaemia), variables not included in previous metaanalyses. We sought studies that included a control group not using PPIs, and excluded studies investigating PPI interaction with clopidogrel. With ischaemic heart disease and stroke causing one in three and one in 20 deaths in the USA annually respectively [20], proton pump inhibitors should be thoroughly investigated for any contribution to atherosclerotic disease.
49
Methods
50
This systematic review was carried out as per PRISMA guidelines [21].
51
Regarding study selection, eligibility criteria are as follows. Studies were included if they examined death or atherosclerotic events (including myocardial infarct, stroke, or peripheral arterial events), and compared a group exposed to proton pump inhibitors with a control group (not exposed to PPIs), in any group of patients. Duration of follow-up was not an exclusion criteria, as the authors were aware the data is limited. Information sources included medical databases Pubmed, Medline, and the Cochrane Review Database. Search strategy involved using the terms ‘proton pump inhibitor’ combined separately with ‘atherosclerosis’, ‘myocardial infarct’, ‘ischemia’, ‘endovascular’, and ‘death’, with the phrase ‘clopidogrel’ excluded. English-language studies from January 1990 to October 2016 were considered, with date last searched 11 November 2016. Attempt was made to contact authors of two of the eligible studies, with response from one. Study designs to be considered included randomised controlled trials, cohort, and case-control studies. Data was abstracted by one reviewer, with a second reviewer assessing for completeness and errors. The data
[1_TD$IF]Figure 1 PRISMA study selection flow diagram. Abbreviation: PPI [2_TD$IF]= proton pump inhibitor. * See text for reasons for exclusion.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
HLC 2521 1–8
3
A Systematic Review
Q1
Table 1 Study characteristics for mortality analysis. Reference
Objective
Study design Population
Comparison
Outcome Years of Data Follow-Up Collection
[3_TD$IF]Maggio (2013)
Effect of PPI on
[9]
all cause mortality comparison
age in Italy
Charlot (2011)
Effect of PPI on
Patients >30 years
[8]
all cause mortality comparison
Prospective Retrospective
Patient >65 years of
PPI vs no PPI All cause [4_TD$IF]2007–?
(days) 365
mortality PPI vs no PPI All cause [5_TD$IF]1997–2006
of age post MI
365
mortality
in Denmark via database Fortuna (2016) Effect of PPI on [30]
Retrospective
all cause mortality comparison
CAD patients >40 years of PPI vs no PPI All cause [6_TD$IF]2007–2009 age using aspirin in
1131.5
mortality
USA via database Abbreviations: PPI [7_TD$IF]= proton pump inhibitor. MI = myocardial infarct. CAD = coronary artery disease.
73
was then extracted into a pre-determined spreadsheet, including author, year of publication, study objective, study design, population demographics, intervention and control groups, outcome, years of data collection, duration of followup, as well as raw numbers of event in both the control and intervention group, and the total number of patients in each group. Risk of bias in each study was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidelines for quality of evidence assessment [22], via the following four domains (ranked low, high or unclear); development and application of eligibility criteria, measurement of exposure and outcome, control of confounding, and completeness of follow-up. The GRADE guidelines were used as the studies were observational rather than randomised trials. All data was dichotomous, and was expressed as odds ratios (OR). Random-effects modelling was used. All analyses were performed using RevMan software, Version 5.3. Funnel plots were not generated due to the small number of included studies. A p value of less than 0.05 was considered to be significant.
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
Q1
Results Systematic review of the literature revealed 37 studies that were considered for inclusion, of which 31 were excluded, leaving six studies directly comparing effect of PPI use on either mortality (three studies), and/or examining the relationship of PPI use with myocardial infarct, stroke, or peripheral arterial event ([1_TD$IF]Figure 1). Reasons for exclusion included: did not examine PPI vs death or cardiac events in humans (23 studies), review or commentary articles (4), examined effect of PPIs on clopidogrel (5), data not amenable for analysis (2). There were no randomised controlled trials, and only one study examined vascular complications (Shah et al. [2]), but this data was not available for analysis. Three studies examined the effect of PPIs on all-cause mortality, and were included in the analysis (Table 1). Total deaths in the PPI-exposed group were 765 of a total of 4,775, while deaths in the non-PPI group was 1,794 out of 17,652. As seen in Table 2, mortality rate was significantly higher in patients taking PPIs (OR 1.68 [95% CI 1.53–1.84), p < 0.001).
Table 2 Forest plot of odds ratios (ORs) of mortality analysis in patients using proton pump inhibitors (PPI) versus those without (control). Squares represent the odds ratio of the individual studies, with lines indicating 95% confidence intervals (CI). Black diamond represents the Total 95% CI.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 Q9 112
HLC 2521 1–8
4
Six datasets from five studies (Shih et al. [28_TD$IF][23] carried out both a propensity-matched study, annotated with a superscript ‘a’, and a case-control study, annotated with a superscript ‘b’) examining major cardiovascular events were included (Table 3). Four datasets examined myocardial infarct, and two examined major cardiac events (including myocardial infarction, stroke). Total events in PPI patients were 3173, of a total of 169,214 PPI patients considered, as compared to 4497 non-PPI patients of 185,232. As demonstrated in Table 4, rate of major cardiovascular events was significantly higher in patients taking PPIs (OR 1.54 [95% CI 1.11–2.13], p = 0.01). The majority of studies had a low risk of bias across all four domains, although only two datasets (from one study) had a low risk of bias regarding eligibility criteria, with control and PPI groups being heterogenous to unclear degrees (Tables 5 and 6 ).
1131.5
720
14
120
766.5
365 [10_TD$IF]1997–2006
116
(days)
115
Follow-Up
114
Collection
113
Years of Data
T.P. Shiraev, A. Bullen
123
[6_TD$IF]2007–2009
122
[14_TD$IF]2011–2016
120 121
2000–2009
119
[12_TD$IF]2000–2009
118
1994–2011
117
160 161 162 163 164 165
MI
MI, PCI
& stroke
PPI vs
no PPI
CAD patients >40 years of age
using aspirin in USA via database
Retrospective
comparison
Patients post-PCI in Pakistan comparison
Retrospective
no PPI Taiwan via database study
PPI vs Patients >18 years of age in Case control
Note that two datasets were considered in the study by Shah [16_TD$IF]et al., and it is the ‘Stride’ dataset that is considered here.
159
*
158
cardiovascular events
156 157
Effect of PPI on
155
Fortuna (2016) [30]
154
cardiac events post PCI
153
Effect of PPI on
152
Ayub (2016) [29]
151
cardiac events
150
Shih (2014b) [23]
148 149
Effect of PPI on
Taiwan via database comparison
147
Effect of PPI on
146
[9_TD$IF]Charlot (2011) [8]
145
design
144
Study
143
Objective
142
Discussion The current study observed a significant increase in mortality and morbidity due to thrombotic complications (namely ischaemic heart disease and strokes) in patients using proton pump inhibitors. There were no randomised controlled trials, and only one study was found examining peripheral vascular outcomes (Meltzer et al. [24]), but was excluded as it considered interactions with clopidogrel. Initial concern regarding cardiovascular morbidity and mortality secondary to PPIs focussed on interactions with the antiplatelet agent clopidogrel. Clopidogrel is a thioenopyridine which binds to the P2Y12 receptor on platelet surfaces, preventing ADP binding, and therefore preventing platelet activation, key in the ‘pathogenesis and propagation of vascular disease’ [25]. Clopidogrel is activated by hepatic metabolism involving the cytochrome P450 enzyme, specifically the CYP2C19 pathway, although this process results in less than 15% of clopidogrel becoming bioactive [26]. Proton pump inhibitors are inactivated via the CYP2C19 pathway [27], and competitive inhibition of this pathway by PPIs has been cited as a reason for the observed reduction in platelet response and increase in cardiovascular morbidity and mortality in patients taking PPIs concomitantly with clopidogrel [25_TD$IF] [3–5]. As such, there were initial recommendations that, for patients taking clopidogrel and at high risk of gastrointestinal bleeding, consideration should be given to either changing the PPI to pantoprazole [28] (which has no effect on CYP2C19 [5], or changing class to the histamine receptor antagonists, which do not interact with clopidogrel [5]. However, a recent review demonstrated that, while in vitro studies have demonstrated significant reduction in platelet inhibition in patients taking clopidogrel and PPIs, in vivo studies have not corroborated this [7]. In fact, patients on concomitant ticagrelor and PPI therapy also suffered higher cardiovascular morbidity, even though ticagrelor is not metabolised by the CYP2C19 pathway [3], suggesting there
Q1
140 141
[8_TD$IF]Table 3 Study characteristics for analysis of cardiovascular events.
139
Reference
138
Population
137
cardiac events
GORD in USA via database Patients >18 years of age in comparison Retrospective
136
Shih (2014a[1_TD$IF]) [23]
Patients >18 years of age with Retrospective
135
cardiac event Effect of PPI on
134
Effect of PPI on
post MI in Denmark via database comparison
132 133
Shah* (2015) [2]
Patients >30 years of age Retrospective
131
cardiovascular events
130
no PPI
128 129
Abbreviations[15_TD$IF]: PPI = proton pump inhibitor, MI = myocardial infarct, GORD = gastro-oesophageal reflux disease, PCI = percutaneous coronary intervention.
Recurrent
MI no PPI PPI vs
PPI vs
MI PPI vs
no PPI
stroke
MI
MI or
no PPI
127
PPI vs
126
Outcome
125
Comparison
124
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
Q10
Q11
HLC 2521 1–8
5
A Systematic Review
Table 4 Forest plot of odds ratios (ORs) of analysis of cardiovascular events in patients using proton pump inhibitors (PPI) versus those without (control). Squares represent the odds ratio of the individual studies, with lines indicating 95% confidence intervals (CI). Black diamond represents the Total 95% CI.
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
may be another mechanism for the observed increase in myocardial reinfarction in patients taking PPIs [2,8,23,29,30]. Recent research regarding the proposed link between PPIs and cardiovascular disease has regenerated interest in the area. Rather than interactions with antiplatelet agents, there may, in fact, be a direct action of PPIs on the endothelium itself, as well as the clotting cascade. Regarding adverse effects on endothelium, Ghebremariam et al. [31] demonstrated that PPIs cause a reduction in both endothelial nitric oxide synthase (eNOS) and inducible nitric oxide synthase (iNOS), key enzymes in nitric oxide production, via inhibition of asymmetrical dimethylarginine (ADMA). Nitric oxide is vaso-protective, as it decreases platelet interactions with the endothelium [32] as well as platelet activation [25]. Further, any reduction in eNOS increases peripheral vascular resistance, and predisposes to inflammation and thrombosis [31], as well as resultant inhibition in NO dependent functions such as angiogenesis [10]. The proposed mechanism for the PPI-induced reduction of NOS is summarised in [1_TD$IF]Figure 2, and is as follows. Proton pump inhibitors have been demonstrated to inhibit activity of dimethylarginine dimethylaminohydrolase (DDAH), the enzyme responsible for metabolism of ADMA. Asymmetrical dimethylarginine is a competitive inhibitor of nitric oxide synthase, and PPIs are known to elevate plasma ADMA levels in both mouse models and humans [31]. The resulting reduction in DDAH
activity and subsequent increase in ADMA activity results in reduction in NO levels, and resultant cardiovascular impairment [31]. There are additional mechanisms by which the NO pathway is affected by PPIs. Proton pump inhibitors have been demonstrated to reduce vitamin B12 absorption, with resultant hyper-homocysteinaemia and subsequent increases in plasma ADMA causing impaired endothelial function [33]. Further, it has been proposed that PPI use results in reduction in nitric oxide creation from dietary nitrate, due to increased gastric pH [34]. The additional mechanism of increased activity of xanthine oxidoreductase (which generates reactive oxygen species) leading to endothelial dysfunction has been proposed [35]. Other molecular mechanisms have been recognised in PPIinduced acceleration of atherosclerotic risk factors. Esomeprazole was demonstrated to cause changes in gene expression including an increase in plasminogen activator inhibitor-1 expression (an inhibitor of fibrinolysis), and induction of telomeric shortening (an indicator of cell senescence) [10]. These esomeprazole-induced changes unrelated to NOS were proposed to be due to impaired endothelial function due to lysosomal alkalinsation, secondary to inhibition of lysosomal proton pump by PPIs [10]. Interestingly, the above effects may be dose-dependent, with higher rates of death in patients on higher dose PPIs [9].
[17_TD$IF]Table 5 Risk of bias summary for studies included for analysis of mortality. Reference
Eligibility Criteria Bias
Measurement of Exposure & Outcome Bias
Confounding Bias
Follow-Up Bias
[3_TD$IF]Maggio (2013) [9]
Unclear
Low
Low
Low
[18_TD$IF]Charlot (2011) [8]
Unclear
Low
Low
Low
[19_TD$IF]Fortuna (2016) [30]
Unclear
Low
Low
Low
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
HLC 2521 1–8
6
T.P. Shiraev, A. Bullen
Table 6 Risk of bias summary for studies included for analysis of cardiovascular event.
[9_TD$IF]Charlot (2011) [8] [20_TD$IF]Shah (2015) [2]
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
Eligibility Criteria Bias
Measurement of Exposure & Outcome Bias
Confounding Bias
Follow-Up Bias
Unclear Unclear
Low Low
Low Low
Low Low
[21_TD$IF]Shih (2014a[13_TD$IF]) [23]
Low
Low
Low
Low
Shih (2014b) [23]
Low
Low
Low
Low
Ayub (2016) [29]
Unclear
Low
Low
Low
Fortuna (2016) [30]
Unclear
Low
Low
Low
Three per cent of Caucasians and 15 to 20% of Asians are poor metabolisers of PPIs [26] due to poor CYP2C19 enzyme activity, causing a doubling of plasma PPI levels [36], risking higher rates of endothelial dysfunction. As mentioned earlier, these patients will also poorly activate clopidogrel, putting these people at very high risk of adverse endothelial events due to low levels of activated clopidogrel and high levels of PPI. There has been a suggestion that the interaction between PPIs and subsequent myocardial event is that PPIs may be prescribed when the physician misinterprets angina symptoms for gastric symptoms [17,18]. Further, it has been noted that PPI users have a higher rate of comorbidity [9], a potential confounding factor in these studies. Of interest, there is some suggestion that PPIs may be involved in cardiac arrhythmias (see Zhu [29_TD$IF]et al. for review [37]). There are several limitations of the current systematic review, both due to the quality of the studies examined, and the inclusion criteria. The included studies were heterogenous, both in terms of study design, outcomes examined,
and proton pump inhibitor used. Most patient populations were different; some examined patients post cardiac event, some all-comers. Numerous studies suffered selection bias, with the PPI and control groups being different in terms of demographics at baseline. Importantly, as aspirin is recommended by current guidelines for prevention of both primary and secondary cardiovascular disease [38], the lack of strict controlling for aspirin intake is a potential limitation of the current studies. However, all studies attempted to adjust for confounders. Also, the prescription of PPIs is no guarantee of compliance with taking the medication. Further, no randomised controlled trials were found, only one study was prospective, and the patient populations were varied. Regarding the systematic review itself, only English language literature was reviewed, potentially missing a large reservoir of relevant research. The current uncertainty leaves us with clinical equipoise regarding patients with GORD or at risk of gastrointestinal bleeding who concomitantly have, or are at risk of vascular disease. We suggest that consideration should be given to
Figure 2 Proposed mechanism for proton pump inhibitor (PPI)-induced reduction in nitric oxide synthase.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
HLC 2521 1–8
7
A Systematic Review
258
275
either changing PPI to pantoprazole, or changing class to the histamine receptor antagonists (the reasons for which are discussed above). Further, the current practice of prescription of PPIs in all patients (whether vasculopath or not) deserves some thought. Firstly regarding necessity of PPIs, secondly regarding duration of prescription (namely whether a short course will suffice), and thirdly the type of anti-acid therapy chosen. However, with the number needed to harm previously found to be 4357 for myocardial infarction [23], PPIs certainly have their place with appropriate caution. In the United States, PPIs are available over the counter, with no medical supervision required, and recent expert consensus suggests that this should be reviewed [39]. Further research is needed in this area, both at a molecular and population level. Only one prospective study has been carried out in the area, with no randomised controlled trials (although the latter has obvious ethical consequences if PPIs are in fact linked to poor outcomes).
276
Conclusion
277
285
In conclusion, the current systematic review demonstrates a significant increase in morbidity due to cardiovascular disease, and all cause mortality in patients using proton pump inhibitors. Recent research has provided the molecular mechanisms for this association, so careful consideration should be given to the prescription of proton pump inhibitors while clinical equipoise remains. However, the studies analysed were heterogenous with numerous cofounders, so our study findings should be interpreted with caution.
286
Acknowledgements
287
Thanks to Chris Thomas for the inspiration for this article. Thanks to Professor Nigam Shah for provision of study data[30_TD$IF].
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
278 279 280 281 282 283 284
288
289
Appendix A. Supplementary data
290
Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.hlc.2017. 10.020.
291 292 293
References
294 295 296 297 298 299 300 301 302 303 304 305
Q12
Q13
[1] Mo C, Sun G, Wang YZ, et al. PPI versus histamine H2 receptor antagonists for prevention of upper gastrointestinal injury associated with low-dose aspirin: systematic review and meta-analysis. PLoS One 2015;10:e0131558. [2] Shah NH, LePendu P, Bauer-Mehren A, et al. Proton pump inhibitor usage and the risk of myocardial infarction in the general population. PLoS One 2015;10:e0124653. [3] Goodman SG, Clare R, Pieper KS, et al. Association of proton pump inhibitor use on cardiovascular outcomes with clopidogrel and ticagrelor: insights from the platelet inhibition and patient outcomes trial. Circulation 2012;125:978–86.
[4] Ho PM, Maddox TM, Wang L, et al. Risk of adverse outcomes 306 associated with concomitant use of clopidogrel and proton pump 307 inhibitors following acute coronary syndrome. JAMA 2009;301: 308 937–44. 309 [5] Juurlink DN, Gomes T, Ko DT, et al. A population-based study of the 310 drug interaction between proton pump inhibitors and clopidogrel. CMAJ 311 2009;180:713–8. 312 [6] FDA. FDA announces new warning on plavix: avoid use with prilosec/ Q14 313 prilosec OTC; 2009. 314 [7] Focks JJ, Brouwer MA, van Oijen MG, et al. Concomitant use of clopi315 dogrel and proton pump inhibitors: impact on platelet function and 316 clinical outcome—a systematic review. Heart 2013;99:520–7. 317 [8] Charlot M, Grove EL, Hansen PR, et al. Proton pump inhibitor use and 318 risk of adverse cardiovascular events in aspirin treated patients with first 319 time myocardial infarction: nationwide propensity score matched study. 320 BMJ 2011;342:d2690. 321 [9] Maggio M, Corsonello A, Ceda GP, et al. Proton pump inhibitors and risk 322 of 1-year mortality and rehospitalization in older patients discharged 323 from acute care hospitals. JAMA Intern Med 2013;173:518–23. 324 [10] Yepuri G, Sukhovershin R, Nazari-Shafti TZ, et al. Proton pump inhib325 itors accelerate endothelial senescence. Circ Res 2016;118:e36–42. 326 [11] Serbin MA, Guzauskas GF, Veenstra DL. Clopidogrel-proton pump 327 inhibitor drug-drug interaction and risk of adverse clinical outcomes 328 among PCI-treated ACS patients: a meta-analysis. J Manag Care Spec 329 Pharm 2016;22:939–47. 330 [12] Sherwood MW, Melloni C, Jones WS, et al. Individual proton pump 331 inhibitors and outcomes in patients with coronary artery disease on dual 332 antiplatelet therapy: a systematic review. J Am Heart Assoc 2015;4. 333 [13] Bundhun PK, Teeluck AR, Bhurtu A, et al. Is the concomitant use of 334 clopidogrel and Proton Pump Inhibitors still associated with increased 335 adverse cardiovascular outcomes following coronary angioplasty?: a 336 systematic review and meta-analysis of recently published studies 337 (2012–2016). BMC Cardiovasc Disord 2017;17:3. 338 [14] Niu Q, Wang Z, Zhang Y, et al. Combination use of clopidogrel and Q15 339 proton pump inhibitors increases major adverse cardiovascular events in 340 patients with coronary artery disease: a meta-Analysis. J Cardiovasc 341 Pharmacol Ther 2016. 342 [15] Sun S, Cui Z, Zhou M, et al. Proton pump inhibitor monotherapy and the 343 risk of cardiovascular events in patients with gastro-esophageal reflux 344 disease: a meta-analysis. Neurogastroenterol Motil 2017;29. 345 [16] Mandurino-Mirizzi A, Leonardi S, Melloni C. Concomitant use of proton 346 pump inhibitors and dual antiplatelet therapy for cardiovascular out347 comes. Minerva Endocrinol 2017;42:228–37. 348 [17] Juurlink DN, Dormuth CR, Huang A, et al. Proton pump inhibitors and 349 the risk of adverse cardiac events. PLoS One 2013;8:e84890. 350 [18] Turkiewicz A, Vicente RP, Ohlsson H, et al. Revising the link between 351 proton-pump inhibitors and risk of acute myocardial infarction-a case352 crossover analysis. Eur J Clin Pharmacol 2015;71:125–9. 353 [19] Vaduganathan M, Cannon CP, Cryer BL, et al. Efficacy and safety of 354 proton-pump inhibitors in high-risk cardiovascular subsets of the 355 COGENT trial. Am J Med 2016;129:1002–5. 356 [20] Go AS, Mozaffarian D, Roger VL, et al. Executive summary: heart disease 357 and stroke statistics—2014 update: a report from the American Heart 358 Association. Circulation 2014;129:399–410. 359 [21] Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for 360 systematic reviews and meta-analyses: the PRISMA statement. BMJ 361 2009;339:b2535. 362 [22] Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating the 363 quality of evidence—study limitations (risk of bias). J Clin Epidemiol 364 2011;64:407–15. 365 [23] Shih CJ, Chen YT, Ou SM, et al. Proton pump inhibitor use represents an 366 independent risk factor for myocardial infarction. Int J Cardiol 367 2014;177:292–7. 368 [24] Meltzer AJ, Da Silva P, Schneider DB, et al. Clinical significance of the 369 clopidogrel-proton pump inhibitor interaction after peripheral endovas370 cular intervention for claudication. Vasc Endovasc Surg 2012;46:524–9. 371 [25] Fitridge R, Thompson M. Mechanisms of vascular disease: a reference 372 book for vascular specialists. Adelaide, Australia: University of Adelaide 373 Press; 2011. 374 [26] Dean L. Clopidogrel therapy and CYP 2C19 genotype. Medical genetics 375 summaries. National Center for Biotechnology Information; 2012. 376 [27] Shin JM, Kim N. Pharmacokinetics and pharmacodynamics of the proton 377 pump inhibitors. J Neurogastroenterol Motil 2013;19:25–35. 378 [28] Drepper MD, Spahr L, Frossard JL. Clopidogrel and proton pump 379 inhibitors—where do we stand in 2012? World J Gastroenterol 380 2012;18:2161–71. 381
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
HLC 2521 1–8
8
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
T.P. Shiraev, A. Bullen
[29] Ayub A, Parkash O, Naeem B, et al. Proton pump inhibitors and other disease-based factors in the recurrence of adverse cardiovascular events following percutaneous coronary angiography: a long-term cohort. Indian J Gastroenterol 2016;35:117–22. [30] Fortuna LA, Pawloski PA, Parker ED, et al. Proton pump inhibitor use by aspirin-treated coronary artery disease patients is not associated with increased risk of cardiovascular events. Eur Heart J Cardiovasc Pharmacother 2016;2:13–9. [31] Ghebremariam YT, LePendu P, Lee JC, et al. Unexpected effect of proton pump inhibitors: elevation of the cardiovascular risk factor asymmetric dimethylarginine. Circulation 2013;128:845–53. [32] Ghebremariam YT, Lee JC, LePendu P, et al. Response to letters regarding article, ‘‘Unexpected effect of proton pump inhibitors: elevation of the cardiovascular risk factor asymmetric dimethylarginine”. Circulation 2014;129:e428. [33] Stuhlinger MC, Oka RK, Graf EE, et al. Endothelial dysfunction induced by hyperhomocyst(e)inemia: role of asymmetric dimethylarginine. Circulation 2003;108:933–8. [34] Montenegro MF, Lundberg JO. Letter by Montenegro and Lundberg regarding article, ‘‘Unexpected effect of proton pump inhibitors:
[35]
[36]
[37] [38]
[39]
elevation of the cardiovascular risk factor asymmetric dimethylarginine”. Circulation 2014;129:e426. Pinheiro LC, Oliveira-Paula GH, Portella RL, et al. Omeprazole impairs vascular redox biology and causes xanthine oxidoreductase-mediated endothelial dysfunction. Redox Biol 2016;9:134–43. Tang HL, Li Y, Hu YF, et al. Effects of CYP2C19 loss-of-function variants on the eradication of H. pylori infection in patients treated with proton pump inhibitor-based triple therapy regimens: a meta-analysis of randomized clinical trials. PLoS One 2013;8:e62162. Zhu W, Hong K. Potential cardiovascular risks of proton pump inhibitors in the general population. Int Heart J 2017;58:163–6. Vandvik PO, Lincoff AM, Gore JM, et al. Primary and secondary prevention of cardiovascular disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141:e637S–68S. Johnson DA, Katz PO, Armstrong D, et al. The safety of appropriate use of over-the-counter proton pump inhibitors: an evidence-based review and Delphi consensus. Drugs 2017;77:547–61.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419
REVIEW
Heart, Lung and Circulation (2017) xx, 1–8 1443-9506/04/$36.00 https://doi.org/10.1016/j.hlc.2017.10.020
Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review
1 2 3 4
[TD$FIRSNAME]Timothy P[2_TD$IF].[TD$FIRSNAME.] [TD$SURNAME]Shiraev[TD$SURNAME.] *[19_TD$IF], [TD$FIRSNAME]Andrew[TD$FIRSNAME.] [TD$SURNAME]Bullen[TD$SURNAME.]
5
Q1
6 7 8
Q2 Department of Vascular Surgery, St George Hospital, Kogarah, NSW, Australia Received 4 April 2017; received in revised form 15 October 2017; accepted 26 October 2017; online published-ahead-of-print xxx
Q4
Background
Proton pump inhibitors (PPIs) are a commonly prescribed medication that recent data has linked to an increased risk of cardiovascular morbidity and all cause morbidity. The current study sought to perform a systematic review to investigate the link between PPIs and morbidity and mortality
Methods
A systematic review was carried out as per the PRISMA guidelines, with information databases including Pubmed, Medline, and the Cochrane Review Database. English-language studies of all types published from January 1990 to October 2016 were considered. Dichotomous analysis generating odds ratios was performed using RevMan Version 5.3.
Results
Thirty-seven studies were considered, of which five directly compared the effect of PPI use on mortality and/or cardiovascular morbidity (including 22,427 patients in mortality datasets, and 354,446 patients in morbidity datasets). For patients taking PPIs, all cause mortality (OR 1.68 [95% CI 1.53–1.84), p [24_TD$IF]< 0.001) and rate of major cardiovascular events (OR 1.54 [95% CI 1.11–2.13], p = 0.01) were significantly higher.
Conclusions
The current systematic review demonstrates that, in patients using PPIs, there was a significant increase in morbidity due to cardiovascular disease. Careful consideration should be given to the prescription of PPIs while clinical equipoise remains. Further research in the area is required.
Q5
Q6
Keywords
Proton pump inhibitor Myocardial infarct Cardiovascular disease AMDA Nitric oxide Mortality
9 10 11
Introduction Q7
12 13 14 15 16 17 18 19 20 21
Q8
Proton pump inhibitors (PPIs) are indicated in gastroesophageal reflux disease (GORD), and are often the first line in anti-acid therapy as they have been demonstrated to be superior to H2 receptor antagonists [1]. Internationally, over 113 million PPI prescriptions are filled annually, accounting for $13 billion in sales [2]. However, PPIs are not free of side effects, with evidence for increased risks of diarrhoea, pneumonia and fractures in long-term users [1]. PPIs are often also prescribed to counteract the gastrointestinal complications of the antiplatelet aspirin, itself given to prevent atherosclerotic disease.
In the past decade, a number of studies have suggested there was a marked increase in cardiovascular morbidity and mortality in patients taking PPIs concomitantly with clopidogrel [25_TD$IF][3–5], to the extent that the FDA issued a formal warning about this interaction [6]. However, this has recently been suggested not to be a true interaction [7], especially as an increase in mortality and cardiac events has been seen even in patients not taking clopidogrel [8,9], and mechanisms for this increased thrombotic risk have recently been proposed [10]. Subsequent meta-analyses agree that there is an increase in adverse cardiac events in patients even not taking clopidogrel [26_TD$IF][11–16] and work by Sherwood et al. suggested that degree of cardiovascular risk
Q3 *Corresponding[23_TD$IF] author. Tel.: +61 9113 1111, Fax: +6191133984. Email: [email protected] Crown Copyright © 2017 Published by Elsevier B.V. on behalf of Australian and New Zealand Society of Cardiac and Thoracic Surgeons (ANZSCTS) and the Cardiac Society of Australia and New Zealand (CSANZ). All rights reserved.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
22 23 24 25 26 27 28 29 30 31 32 33
HLC 2521 1–8
2
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
T.P. Shiraev, A. Bullen
even differs between individual PPIs [12]. Debate remains as to whether PPIs do truly carry an increase in cardiovascular morbidity and mortality [27_TD$IF][17–19], and to what extent. We sought to investigate the impact of proton pump inhibitor use on both mortality, and morbidity due to thrombotic complications, namely ischaemic heart disease and vascular disease (including cerebrovascular accidents and limb ischaemia), variables not included in previous metaanalyses. We sought studies that included a control group not using PPIs, and excluded studies investigating PPI interaction with clopidogrel. With ischaemic heart disease and stroke causing one in three and one in 20 deaths in the USA annually respectively [20], proton pump inhibitors should be thoroughly investigated for any contribution to atherosclerotic disease.
49
Methods
50
This systematic review was carried out as per PRISMA guidelines [21].
51
Regarding study selection, eligibility criteria are as follows. Studies were included if they examined death or atherosclerotic events (including myocardial infarct, stroke, or peripheral arterial events), and compared a group exposed to proton pump inhibitors with a control group (not exposed to PPIs), in any group of patients. Duration of follow-up was not an exclusion criteria, as the authors were aware the data is limited. Information sources included medical databases Pubmed, Medline, and the Cochrane Review Database. Search strategy involved using the terms ‘proton pump inhibitor’ combined separately with ‘atherosclerosis’, ‘myocardial infarct’, ‘ischemia’, ‘endovascular’, and ‘death’, with the phrase ‘clopidogrel’ excluded. English-language studies from January 1990 to October 2016 were considered, with date last searched 11 November 2016. Attempt was made to contact authors of two of the eligible studies, with response from one. Study designs to be considered included randomised controlled trials, cohort, and case-control studies. Data was abstracted by one reviewer, with a second reviewer assessing for completeness and errors. The data
[1_TD$IF]Figure 1 PRISMA study selection flow diagram. Abbreviation: PPI [2_TD$IF]= proton pump inhibitor. * See text for reasons for exclusion.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72
HLC 2521 1–8
3
A Systematic Review
Q1
Table 1 Study characteristics for mortality analysis. Reference
Objective
Study design Population
Comparison
Outcome Years of Data Follow-Up Collection
[3_TD$IF]Maggio (2013)
Effect of PPI on
[9]
all cause mortality comparison
age in Italy
Charlot (2011)
Effect of PPI on
Patients >30 years
[8]
all cause mortality comparison
Prospective Retrospective
Patient >65 years of
PPI vs no PPI All cause [4_TD$IF]2007–?
(days) 365
mortality PPI vs no PPI All cause [5_TD$IF]1997–2006
of age post MI
365
mortality
in Denmark via database Fortuna (2016) Effect of PPI on [30]
Retrospective
all cause mortality comparison
CAD patients >40 years of PPI vs no PPI All cause [6_TD$IF]2007–2009 age using aspirin in
1131.5
mortality
USA via database Abbreviations: PPI [7_TD$IF]= proton pump inhibitor. MI = myocardial infarct. CAD = coronary artery disease.
73
was then extracted into a pre-determined spreadsheet, including author, year of publication, study objective, study design, population demographics, intervention and control groups, outcome, years of data collection, duration of followup, as well as raw numbers of event in both the control and intervention group, and the total number of patients in each group. Risk of bias in each study was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) guidelines for quality of evidence assessment [22], via the following four domains (ranked low, high or unclear); development and application of eligibility criteria, measurement of exposure and outcome, control of confounding, and completeness of follow-up. The GRADE guidelines were used as the studies were observational rather than randomised trials. All data was dichotomous, and was expressed as odds ratios (OR). Random-effects modelling was used. All analyses were performed using RevMan software, Version 5.3. Funnel plots were not generated due to the small number of included studies. A p value of less than 0.05 was considered to be significant.
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
Q1
Results Systematic review of the literature revealed 37 studies that were considered for inclusion, of which 31 were excluded, leaving six studies directly comparing effect of PPI use on either mortality (three studies), and/or examining the relationship of PPI use with myocardial infarct, stroke, or peripheral arterial event ([1_TD$IF]Figure 1). Reasons for exclusion included: did not examine PPI vs death or cardiac events in humans (23 studies), review or commentary articles (4), examined effect of PPIs on clopidogrel (5), data not amenable for analysis (2). There were no randomised controlled trials, and only one study examined vascular complications (Shah et al. [2]), but this data was not available for analysis. Three studies examined the effect of PPIs on all-cause mortality, and were included in the analysis (Table 1). Total deaths in the PPI-exposed group were 765 of a total of 4,775, while deaths in the non-PPI group was 1,794 out of 17,652. As seen in Table 2, mortality rate was significantly higher in patients taking PPIs (OR 1.68 [95% CI 1.53–1.84), p < 0.001).
Table 2 Forest plot of odds ratios (ORs) of mortality analysis in patients using proton pump inhibitors (PPI) versus those without (control). Squares represent the odds ratio of the individual studies, with lines indicating 95% confidence intervals (CI). Black diamond represents the Total 95% CI.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 Q9 112
HLC 2521 1–8
4
Six datasets from five studies (Shih et al. [28_TD$IF][23] carried out both a propensity-matched study, annotated with a superscript ‘a’, and a case-control study, annotated with a superscript ‘b’) examining major cardiovascular events were included (Table 3). Four datasets examined myocardial infarct, and two examined major cardiac events (including myocardial infarction, stroke). Total events in PPI patients were 3173, of a total of 169,214 PPI patients considered, as compared to 4497 non-PPI patients of 185,232. As demonstrated in Table 4, rate of major cardiovascular events was significantly higher in patients taking PPIs (OR 1.54 [95% CI 1.11–2.13], p = 0.01). The majority of studies had a low risk of bias across all four domains, although only two datasets (from one study) had a low risk of bias regarding eligibility criteria, with control and PPI groups being heterogenous to unclear degrees (Tables 5 and 6 ).
1131.5
720
14
120
766.5
365 [10_TD$IF]1997–2006
116
(days)
115
Follow-Up
114
Collection
113
Years of Data
T.P. Shiraev, A. Bullen
123
[6_TD$IF]2007–2009
122
[14_TD$IF]2011–2016
120 121
2000–2009
119
[12_TD$IF]2000–2009
118
1994–2011
117
160 161 162 163 164 165
MI
MI, PCI
& stroke
PPI vs
no PPI
CAD patients >40 years of age
using aspirin in USA via database
Retrospective
comparison
Patients post-PCI in Pakistan comparison
Retrospective
no PPI Taiwan via database study
PPI vs Patients >18 years of age in Case control
Note that two datasets were considered in the study by Shah [16_TD$IF]et al., and it is the ‘Stride’ dataset that is considered here.
159
*
158
cardiovascular events
156 157
Effect of PPI on
155
Fortuna (2016) [30]
154
cardiac events post PCI
153
Effect of PPI on
152
Ayub (2016) [29]
151
cardiac events
150
Shih (2014b) [23]
148 149
Effect of PPI on
Taiwan via database comparison
147
Effect of PPI on
146
[9_TD$IF]Charlot (2011) [8]
145
design
144
Study
143
Objective
142
Discussion The current study observed a significant increase in mortality and morbidity due to thrombotic complications (namely ischaemic heart disease and strokes) in patients using proton pump inhibitors. There were no randomised controlled trials, and only one study was found examining peripheral vascular outcomes (Meltzer et al. [24]), but was excluded as it considered interactions with clopidogrel. Initial concern regarding cardiovascular morbidity and mortality secondary to PPIs focussed on interactions with the antiplatelet agent clopidogrel. Clopidogrel is a thioenopyridine which binds to the P2Y12 receptor on platelet surfaces, preventing ADP binding, and therefore preventing platelet activation, key in the ‘pathogenesis and propagation of vascular disease’ [25]. Clopidogrel is activated by hepatic metabolism involving the cytochrome P450 enzyme, specifically the CYP2C19 pathway, although this process results in less than 15% of clopidogrel becoming bioactive [26]. Proton pump inhibitors are inactivated via the CYP2C19 pathway [27], and competitive inhibition of this pathway by PPIs has been cited as a reason for the observed reduction in platelet response and increase in cardiovascular morbidity and mortality in patients taking PPIs concomitantly with clopidogrel [25_TD$IF] [3–5]. As such, there were initial recommendations that, for patients taking clopidogrel and at high risk of gastrointestinal bleeding, consideration should be given to either changing the PPI to pantoprazole [28] (which has no effect on CYP2C19 [5], or changing class to the histamine receptor antagonists, which do not interact with clopidogrel [5]. However, a recent review demonstrated that, while in vitro studies have demonstrated significant reduction in platelet inhibition in patients taking clopidogrel and PPIs, in vivo studies have not corroborated this [7]. In fact, patients on concomitant ticagrelor and PPI therapy also suffered higher cardiovascular morbidity, even though ticagrelor is not metabolised by the CYP2C19 pathway [3], suggesting there
Q1
140 141
[8_TD$IF]Table 3 Study characteristics for analysis of cardiovascular events.
139
Reference
138
Population
137
cardiac events
GORD in USA via database Patients >18 years of age in comparison Retrospective
136
Shih (2014a[1_TD$IF]) [23]
Patients >18 years of age with Retrospective
135
cardiac event Effect of PPI on
134
Effect of PPI on
post MI in Denmark via database comparison
132 133
Shah* (2015) [2]
Patients >30 years of age Retrospective
131
cardiovascular events
130
no PPI
128 129
Abbreviations[15_TD$IF]: PPI = proton pump inhibitor, MI = myocardial infarct, GORD = gastro-oesophageal reflux disease, PCI = percutaneous coronary intervention.
Recurrent
MI no PPI PPI vs
PPI vs
MI PPI vs
no PPI
stroke
MI
MI or
no PPI
127
PPI vs
126
Outcome
125
Comparison
124
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
Q10
Q11
HLC 2521 1–8
5
A Systematic Review
Table 4 Forest plot of odds ratios (ORs) of analysis of cardiovascular events in patients using proton pump inhibitors (PPI) versus those without (control). Squares represent the odds ratio of the individual studies, with lines indicating 95% confidence intervals (CI). Black diamond represents the Total 95% CI.
166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
may be another mechanism for the observed increase in myocardial reinfarction in patients taking PPIs [2,8,23,29,30]. Recent research regarding the proposed link between PPIs and cardiovascular disease has regenerated interest in the area. Rather than interactions with antiplatelet agents, there may, in fact, be a direct action of PPIs on the endothelium itself, as well as the clotting cascade. Regarding adverse effects on endothelium, Ghebremariam et al. [31] demonstrated that PPIs cause a reduction in both endothelial nitric oxide synthase (eNOS) and inducible nitric oxide synthase (iNOS), key enzymes in nitric oxide production, via inhibition of asymmetrical dimethylarginine (ADMA). Nitric oxide is vaso-protective, as it decreases platelet interactions with the endothelium [32] as well as platelet activation [25]. Further, any reduction in eNOS increases peripheral vascular resistance, and predisposes to inflammation and thrombosis [31], as well as resultant inhibition in NO dependent functions such as angiogenesis [10]. The proposed mechanism for the PPI-induced reduction of NOS is summarised in [1_TD$IF]Figure 2, and is as follows. Proton pump inhibitors have been demonstrated to inhibit activity of dimethylarginine dimethylaminohydrolase (DDAH), the enzyme responsible for metabolism of ADMA. Asymmetrical dimethylarginine is a competitive inhibitor of nitric oxide synthase, and PPIs are known to elevate plasma ADMA levels in both mouse models and humans [31]. The resulting reduction in DDAH
activity and subsequent increase in ADMA activity results in reduction in NO levels, and resultant cardiovascular impairment [31]. There are additional mechanisms by which the NO pathway is affected by PPIs. Proton pump inhibitors have been demonstrated to reduce vitamin B12 absorption, with resultant hyper-homocysteinaemia and subsequent increases in plasma ADMA causing impaired endothelial function [33]. Further, it has been proposed that PPI use results in reduction in nitric oxide creation from dietary nitrate, due to increased gastric pH [34]. The additional mechanism of increased activity of xanthine oxidoreductase (which generates reactive oxygen species) leading to endothelial dysfunction has been proposed [35]. Other molecular mechanisms have been recognised in PPIinduced acceleration of atherosclerotic risk factors. Esomeprazole was demonstrated to cause changes in gene expression including an increase in plasminogen activator inhibitor-1 expression (an inhibitor of fibrinolysis), and induction of telomeric shortening (an indicator of cell senescence) [10]. These esomeprazole-induced changes unrelated to NOS were proposed to be due to impaired endothelial function due to lysosomal alkalinsation, secondary to inhibition of lysosomal proton pump by PPIs [10]. Interestingly, the above effects may be dose-dependent, with higher rates of death in patients on higher dose PPIs [9].
[17_TD$IF]Table 5 Risk of bias summary for studies included for analysis of mortality. Reference
Eligibility Criteria Bias
Measurement of Exposure & Outcome Bias
Confounding Bias
Follow-Up Bias
[3_TD$IF]Maggio (2013) [9]
Unclear
Low
Low
Low
[18_TD$IF]Charlot (2011) [8]
Unclear
Low
Low
Low
[19_TD$IF]Fortuna (2016) [30]
Unclear
Low
Low
Low
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217
HLC 2521 1–8
6
T.P. Shiraev, A. Bullen
Table 6 Risk of bias summary for studies included for analysis of cardiovascular event.
[9_TD$IF]Charlot (2011) [8] [20_TD$IF]Shah (2015) [2]
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
Eligibility Criteria Bias
Measurement of Exposure & Outcome Bias
Confounding Bias
Follow-Up Bias
Unclear Unclear
Low Low
Low Low
Low Low
[21_TD$IF]Shih (2014a[13_TD$IF]) [23]
Low
Low
Low
Low
Shih (2014b) [23]
Low
Low
Low
Low
Ayub (2016) [29]
Unclear
Low
Low
Low
Fortuna (2016) [30]
Unclear
Low
Low
Low
Three per cent of Caucasians and 15 to 20% of Asians are poor metabolisers of PPIs [26] due to poor CYP2C19 enzyme activity, causing a doubling of plasma PPI levels [36], risking higher rates of endothelial dysfunction. As mentioned earlier, these patients will also poorly activate clopidogrel, putting these people at very high risk of adverse endothelial events due to low levels of activated clopidogrel and high levels of PPI. There has been a suggestion that the interaction between PPIs and subsequent myocardial event is that PPIs may be prescribed when the physician misinterprets angina symptoms for gastric symptoms [17,18]. Further, it has been noted that PPI users have a higher rate of comorbidity [9], a potential confounding factor in these studies. Of interest, there is some suggestion that PPIs may be involved in cardiac arrhythmias (see Zhu [29_TD$IF]et al. for review [37]). There are several limitations of the current systematic review, both due to the quality of the studies examined, and the inclusion criteria. The included studies were heterogenous, both in terms of study design, outcomes examined,
and proton pump inhibitor used. Most patient populations were different; some examined patients post cardiac event, some all-comers. Numerous studies suffered selection bias, with the PPI and control groups being different in terms of demographics at baseline. Importantly, as aspirin is recommended by current guidelines for prevention of both primary and secondary cardiovascular disease [38], the lack of strict controlling for aspirin intake is a potential limitation of the current studies. However, all studies attempted to adjust for confounders. Also, the prescription of PPIs is no guarantee of compliance with taking the medication. Further, no randomised controlled trials were found, only one study was prospective, and the patient populations were varied. Regarding the systematic review itself, only English language literature was reviewed, potentially missing a large reservoir of relevant research. The current uncertainty leaves us with clinical equipoise regarding patients with GORD or at risk of gastrointestinal bleeding who concomitantly have, or are at risk of vascular disease. We suggest that consideration should be given to
Figure 2 Proposed mechanism for proton pump inhibitor (PPI)-induced reduction in nitric oxide synthase.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257
HLC 2521 1–8
7
A Systematic Review
258
275
either changing PPI to pantoprazole, or changing class to the histamine receptor antagonists (the reasons for which are discussed above). Further, the current practice of prescription of PPIs in all patients (whether vasculopath or not) deserves some thought. Firstly regarding necessity of PPIs, secondly regarding duration of prescription (namely whether a short course will suffice), and thirdly the type of anti-acid therapy chosen. However, with the number needed to harm previously found to be 4357 for myocardial infarction [23], PPIs certainly have their place with appropriate caution. In the United States, PPIs are available over the counter, with no medical supervision required, and recent expert consensus suggests that this should be reviewed [39]. Further research is needed in this area, both at a molecular and population level. Only one prospective study has been carried out in the area, with no randomised controlled trials (although the latter has obvious ethical consequences if PPIs are in fact linked to poor outcomes).
276
Conclusion
277
285
In conclusion, the current systematic review demonstrates a significant increase in morbidity due to cardiovascular disease, and all cause mortality in patients using proton pump inhibitors. Recent research has provided the molecular mechanisms for this association, so careful consideration should be given to the prescription of proton pump inhibitors while clinical equipoise remains. However, the studies analysed were heterogenous with numerous cofounders, so our study findings should be interpreted with caution.
286
Acknowledgements
287
Thanks to Chris Thomas for the inspiration for this article. Thanks to Professor Nigam Shah for provision of study data[30_TD$IF].
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
278 279 280 281 282 283 284
288
289
Appendix A. Supplementary data
290
Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.hlc.2017. 10.020.
291 292 293
References
294 295 296 297 298 299 300 301 302 303 304 305
Q12
Q13
[1] Mo C, Sun G, Wang YZ, et al. PPI versus histamine H2 receptor antagonists for prevention of upper gastrointestinal injury associated with low-dose aspirin: systematic review and meta-analysis. PLoS One 2015;10:e0131558. [2] Shah NH, LePendu P, Bauer-Mehren A, et al. Proton pump inhibitor usage and the risk of myocardial infarction in the general population. PLoS One 2015;10:e0124653. [3] Goodman SG, Clare R, Pieper KS, et al. Association of proton pump inhibitor use on cardiovascular outcomes with clopidogrel and ticagrelor: insights from the platelet inhibition and patient outcomes trial. Circulation 2012;125:978–86.
[4] Ho PM, Maddox TM, Wang L, et al. Risk of adverse outcomes 306 associated with concomitant use of clopidogrel and proton pump 307 inhibitors following acute coronary syndrome. JAMA 2009;301: 308 937–44. 309 [5] Juurlink DN, Gomes T, Ko DT, et al. A population-based study of the 310 drug interaction between proton pump inhibitors and clopidogrel. CMAJ 311 2009;180:713–8. 312 [6] FDA. FDA announces new warning on plavix: avoid use with prilosec/ Q14 313 prilosec OTC; 2009. 314 [7] Focks JJ, Brouwer MA, van Oijen MG, et al. Concomitant use of clopi315 dogrel and proton pump inhibitors: impact on platelet function and 316 clinical outcome—a systematic review. Heart 2013;99:520–7. 317 [8] Charlot M, Grove EL, Hansen PR, et al. Proton pump inhibitor use and 318 risk of adverse cardiovascular events in aspirin treated patients with first 319 time myocardial infarction: nationwide propensity score matched study. 320 BMJ 2011;342:d2690. 321 [9] Maggio M, Corsonello A, Ceda GP, et al. Proton pump inhibitors and risk 322 of 1-year mortality and rehospitalization in older patients discharged 323 from acute care hospitals. JAMA Intern Med 2013;173:518–23. 324 [10] Yepuri G, Sukhovershin R, Nazari-Shafti TZ, et al. Proton pump inhib325 itors accelerate endothelial senescence. Circ Res 2016;118:e36–42. 326 [11] Serbin MA, Guzauskas GF, Veenstra DL. Clopidogrel-proton pump 327 inhibitor drug-drug interaction and risk of adverse clinical outcomes 328 among PCI-treated ACS patients: a meta-analysis. J Manag Care Spec 329 Pharm 2016;22:939–47. 330 [12] Sherwood MW, Melloni C, Jones WS, et al. Individual proton pump 331 inhibitors and outcomes in patients with coronary artery disease on dual 332 antiplatelet therapy: a systematic review. J Am Heart Assoc 2015;4. 333 [13] Bundhun PK, Teeluck AR, Bhurtu A, et al. Is the concomitant use of 334 clopidogrel and Proton Pump Inhibitors still associated with increased 335 adverse cardiovascular outcomes following coronary angioplasty?: a 336 systematic review and meta-analysis of recently published studies 337 (2012–2016). BMC Cardiovasc Disord 2017;17:3. 338 [14] Niu Q, Wang Z, Zhang Y, et al. Combination use of clopidogrel and Q15 339 proton pump inhibitors increases major adverse cardiovascular events in 340 patients with coronary artery disease: a meta-Analysis. J Cardiovasc 341 Pharmacol Ther 2016. 342 [15] Sun S, Cui Z, Zhou M, et al. Proton pump inhibitor monotherapy and the 343 risk of cardiovascular events in patients with gastro-esophageal reflux 344 disease: a meta-analysis. Neurogastroenterol Motil 2017;29. 345 [16] Mandurino-Mirizzi A, Leonardi S, Melloni C. Concomitant use of proton 346 pump inhibitors and dual antiplatelet therapy for cardiovascular out347 comes. Minerva Endocrinol 2017;42:228–37. 348 [17] Juurlink DN, Dormuth CR, Huang A, et al. Proton pump inhibitors and 349 the risk of adverse cardiac events. PLoS One 2013;8:e84890. 350 [18] Turkiewicz A, Vicente RP, Ohlsson H, et al. Revising the link between 351 proton-pump inhibitors and risk of acute myocardial infarction-a case352 crossover analysis. Eur J Clin Pharmacol 2015;71:125–9. 353 [19] Vaduganathan M, Cannon CP, Cryer BL, et al. Efficacy and safety of 354 proton-pump inhibitors in high-risk cardiovascular subsets of the 355 COGENT trial. Am J Med 2016;129:1002–5. 356 [20] Go AS, Mozaffarian D, Roger VL, et al. Executive summary: heart disease 357 and stroke statistics—2014 update: a report from the American Heart 358 Association. Circulation 2014;129:399–410. 359 [21] Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for 360 systematic reviews and meta-analyses: the PRISMA statement. BMJ 361 2009;339:b2535. 362 [22] Guyatt GH, Oxman AD, Vist G, et al. GRADE guidelines: 4. Rating the 363 quality of evidence—study limitations (risk of bias). J Clin Epidemiol 364 2011;64:407–15. 365 [23] Shih CJ, Chen YT, Ou SM, et al. Proton pump inhibitor use represents an 366 independent risk factor for myocardial infarction. Int J Cardiol 367 2014;177:292–7. 368 [24] Meltzer AJ, Da Silva P, Schneider DB, et al. Clinical significance of the 369 clopidogrel-proton pump inhibitor interaction after peripheral endovas370 cular intervention for claudication. Vasc Endovasc Surg 2012;46:524–9. 371 [25] Fitridge R, Thompson M. Mechanisms of vascular disease: a reference 372 book for vascular specialists. Adelaide, Australia: University of Adelaide 373 Press; 2011. 374 [26] Dean L. Clopidogrel therapy and CYP 2C19 genotype. Medical genetics 375 summaries. National Center for Biotechnology Information; 2012. 376 [27] Shin JM, Kim N. Pharmacokinetics and pharmacodynamics of the proton 377 pump inhibitors. J Neurogastroenterol Motil 2013;19:25–35. 378 [28] Drepper MD, Spahr L, Frossard JL. Clopidogrel and proton pump 379 inhibitors—where do we stand in 2012? World J Gastroenterol 380 2012;18:2161–71. 381
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
HLC 2521 1–8
8
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
T.P. Shiraev, A. Bullen
[29] Ayub A, Parkash O, Naeem B, et al. Proton pump inhibitors and other disease-based factors in the recurrence of adverse cardiovascular events following percutaneous coronary angiography: a long-term cohort. Indian J Gastroenterol 2016;35:117–22. [30] Fortuna LA, Pawloski PA, Parker ED, et al. Proton pump inhibitor use by aspirin-treated coronary artery disease patients is not associated with increased risk of cardiovascular events. Eur Heart J Cardiovasc Pharmacother 2016;2:13–9. [31] Ghebremariam YT, LePendu P, Lee JC, et al. Unexpected effect of proton pump inhibitors: elevation of the cardiovascular risk factor asymmetric dimethylarginine. Circulation 2013;128:845–53. [32] Ghebremariam YT, Lee JC, LePendu P, et al. Response to letters regarding article, ‘‘Unexpected effect of proton pump inhibitors: elevation of the cardiovascular risk factor asymmetric dimethylarginine”. Circulation 2014;129:e428. [33] Stuhlinger MC, Oka RK, Graf EE, et al. Endothelial dysfunction induced by hyperhomocyst(e)inemia: role of asymmetric dimethylarginine. Circulation 2003;108:933–8. [34] Montenegro MF, Lundberg JO. Letter by Montenegro and Lundberg regarding article, ‘‘Unexpected effect of proton pump inhibitors:
[35]
[36]
[37] [38]
[39]
elevation of the cardiovascular risk factor asymmetric dimethylarginine”. Circulation 2014;129:e426. Pinheiro LC, Oliveira-Paula GH, Portella RL, et al. Omeprazole impairs vascular redox biology and causes xanthine oxidoreductase-mediated endothelial dysfunction. Redox Biol 2016;9:134–43. Tang HL, Li Y, Hu YF, et al. Effects of CYP2C19 loss-of-function variants on the eradication of H. pylori infection in patients treated with proton pump inhibitor-based triple therapy regimens: a meta-analysis of randomized clinical trials. PLoS One 2013;8:e62162. Zhu W, Hong K. Potential cardiovascular risks of proton pump inhibitors in the general population. Int Heart J 2017;58:163–6. Vandvik PO, Lincoff AM, Gore JM, et al. Primary and secondary prevention of cardiovascular disease: Antithrombotic Therapy and Prevention of Thrombosis, 9th ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012;141:e637S–68S. Johnson DA, Katz PO, Armstrong D, et al. The safety of appropriate use of over-the-counter proton pump inhibitors: an evidence-based review and Delphi consensus. Drugs 2017;77:547–61.
Please cite this article in press as: Shiraev TP, Bullen A. Proton Pump Inhibitors and Cardiovascular Events: A Systematic Review. Heart, Lung and Circulation (2017), https://doi.org/10.1016/j.hlc.2017.10.020
402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419