BMJ
Confidential: For Review Only
Angiotensin Converting Enzyme Inhibitors and the Risk of Lung Cancer: population-based cohort study
Journal: BMJ
Manuscript ID BMJ.2018.044369
Article Type: Research
BMJ Journal: BMJ
Date Submitted by the Author: 27-Mar-2018
Complete List of Authors: Hicks, Blánaid; Jewish General Hospital, Centre for Clinical Epidemiology, Lady Davis Institute; McGill University, Department of Epidemiology, Biostatistics and Occupational Health Filion, Kristian; McGill University, Centre for Clinical Epidemiology Yin, Hui; Jewish General Hospital, Centre for Clinical Epidemiology Sakr, Lama; Jewish General Hospital, Division of Pulmonary Diseases, Department of Medicine Udell, Jacob; Women’s College Hospital, Women’s College Research Institute, University of Toronto, Cardiovascular Division, Department of Medicine Azoulay, Laurent; McGill University, Department of Oncology
Keywords: ACE inhibitors, Lung cancer, Population-based cohort
https://mc.manuscriptcentral.com/bmj Page 1 of 42 BMJ
1 2 Angiotensin Converting Enzyme Inhibitors and the Risk of Lung 3 4 5 Cancer: population-based cohort study 6 7 8 Blánaid M. Hicks PhD Postdoctoral Research Fellow1,2,3, Kristian B. Filion PhD Assistant Professor of 9 10 epidemiology1,2,4, Hui Yin MSc Statistiscian1, Lama Sakr MD MSc Pulmonologist 5, 11 Confidential: For Review Only 6,7 12 Jacob A. Udell MD MPH Cardiologist and Assistant Professor of cardiology , Laurent Azoulay PhD 13 1,2,8 14 Associate Professor of epidemiology and oncology 15 16 17 1 Centre for Clinical Epidemiology, Lady Davis Institute, Jewish General Hospital, Montreal, Canada 18 2 19 Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Montreal, 20 Canada 21 22 3 Centre for Public Health, School of Medicine, Dentistry and Biomedical Sciences, Queen's University 23 24 Belfast, Belfast, United Kingdom 25 4 26 Division of Clinical Epidemiology, Department of Medicine, McGill University, Montreal, Canada 27 5 Division of Pulmonary Diseases, Department of Medicine, Jewish General Hospital, Canada 28 29 6 Women's College Research Institute and Cardiovascular Division, Department of Medicine, Women's 30 31 College Hospital, University of Toronto, Canada 32 7 33 Cardiovascular Division, Department of Medicine, Peter Munk Cardiac Centre, Toronto General 34 Hospital, University of Toronto, Canada 35 36 8 Gerald Bronfman Department of Oncology, McGill University, Montreal, Canada 37 38 39 Word count: 3208 40 41 Running Head: ACE inhibitors and lung cancer 42 43 44 45 Correspondence to: 46 Dr Laurent Azoulay 47 Center for Clinical Epidemiology, Lady Davis Institute, 48 49 Jewish General Hospital, 50 3755 Côte Sainte Catherine, H 425.1 51 Montreal, Quebec, Canada, H3T 1E2 52 Tel: 514.340.8222 ext. 28396; Fax: 514.340.7564 53 Email: [email protected] 54 55 56 March 27, 2018 57 58 59 1 60 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 42
1 2 3 ABSTRACT 4 5 Objective: To determine whether the use of angiotensin converting enzyme inhibitors (ACEIs), 6 7 when compared with use of angiotensin receptor blockers (ARBs), is associated with an 8 9 10 increased risk of lung cancer. 11 Confidential: For Review Only 12 Design: Population based cohort study. 13 14 Setting: United Kingdom Clinical Practice Research Datalink. 15 16 17 Participants: A cohort of 992,061 patients newly treated with antihypertensive drugs between 18 19 January 1, 1995 and December 31, 2015 was identified and followed until December 31, 2016. 20 21 Main outcomes and measures: Cox proportional hazards models were used to estimate adjusted 22 23 24 hazard ratios (HRs) with 95% confidence intervals (CIs) of incident lung cancer associated with 25 26 the time varying use of ACEIs, when compared with use of ARBs, overall, by cumulative 27 28 duration of use, and by time since initiation. 29 30 Results: The cohort was followed for a mean of 6.4 years (standard deviation 4.7), generating 31 32 33 7952 incident lung cancer events (crude incidence 1.3 (95% CI 1.2 to 1.3) per 1000 person 34 35 years). Overall, use of ACEIs was associated with an increased risk of lung cancer (incidence 36 37 rate: 1.6 vs 1.2 per 1000 person years; HR: 1.13, 95% CI: 1.00 to 1.29), when compared with use 38 39 40 of ARBs. HRs gradually increased with longer durations of use, with an association evident after 41 42 5 years of use (HR, 1.21; 95% CI, 1.05 to 1.39) and peaking after more than 10 years of use (HR 43 44 1.31, 95% CI: 1.08 to 1.58). Similar findings were observed with time since initiation. 45 46 47 Conclusions: In this population based cohort study, the use of ACEIs was associated with an 48 49 increased risk of lung cancer. The association was particularly elevated among those using 50 51 ACEIs for over 5 years. Although further studies are imperative, clinicians should be aware of 52 53 this potential safety issue. 54 55 56 57 58 59 2 60 https://mc.manuscriptcentral.com/bmj Page 3 of 42 BMJ
1 2 3 INTRODUCTION 4 5 6 Angiotensin converting enzyme inhibitors (ACEIs) are effective drugs used in the 7 8 treatment of hypertension.[1] While these drugs have been shown to be relatively safe in the 9 10 short term, there have been concerns that their long term use may be associated with an 11 Confidential: For Review Only 12 increased risk of cancer. These concerns have been subject to debate, with observational studies 13 14 15 producing mixed findings,[2–4] including with respect to lung cancer.[2,4] Indeed, there is some 16 17 biological evidence for a possible association between ACEIs and lung cancer risk. The use of 18 19 ACEIs causes an accumulation of bradykinin in the lung,[5] which has been reported to stimulate 20 21 22 growth of lung cancer.[5,6] Moreover, ACEI use also results in accumulation of substance P, 23 24 which is expressed in lung cancer tissue and has been associated with tumor proliferation and 25 26 angiogenesis.[7] 27 28 29 While recent meta analyses of randomized controlled trials (RCTs) found no evidence of 30 31 an increase in cancer incidence with ACEIs, most had small sample sizes and short durations of 32 33 follow up (median 3.5 years).[8,9] To date, the few observational studies that have investigated 34 35 the association between ACEI use and lung cancer reported mixed findings.[10–17] However, 36 37 38 most of these studies were designed to assess the risk of cancer overall, and not lung cancer 39 40 specifically.[10–16] Additionally, several of these studies had a number of methodological 41 42 shortcomings, including a short duration of follow up (e.g., median of 0.7 years),[17] failure to 43 44 45 account for cancer latency,[12,13,15,17] and immortal time bias.[15] Furthermore, results of 46 47 some studies may have been influenced by the use of an inappropriate comparator group, 48 49 introducing potential confounding by indication,[14] and the inclusion of prevalent users of 50 51 52 antihypertensives.[15] 53 54 55 56 57 58 59 3 60 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 42
1 2 3 Thus, to address the conflicting and limited evidence from both pre clinical and 4 5 6 observational studies, we conducted a large, population based study to determine whether the use 7 8 of ACEIs, compared with use of angiotensin receptor blockers (ARBs), is associated with an 9 10 increased risk of lung cancer. 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 4 60 https://mc.manuscriptcentral.com/bmj Page 5 of 42 BMJ
1 2 3 METHODS 4 5 6 Data source 7 8 This study was conducted using the United Kingdom (UK) Clinical Practice Research 9 10 Datalink (CPRD). The CPRD includes data from approximately 700 general practices 11 Confidential: For Review Only 12 comprising over 15 million patients; these have been shown to be representative of the UK 13 14 11 15 population. The CPRD records demographic information, anthropometric data (such as body 16 17 mass index [BMI]), lifestyle information (such as smoking status and alcohol use), medical 18 19 diagnoses and procedures (coded using the Read code classification[19]), and prescription data 20 21 22 (coded according to the UK Prescription Pricing Authority Dictionary[20]) which have been 23 24 shown to be valid and of high quality.12,13 Furthermore, lung cancer diagnoses recorded in the 25 26 CPRD have been shown to be highly concordant (>93%) with those recorded in the UK National 27 28 14 29 Cancer Data Repository. 30 31 The study protocol was approved by the Independent Scientific Advisory Committee of 32 33 the CPRD (protocol number 16_255R) and by the Research Ethics Board of the Jewish General 34 35 Hospital, Montreal, Canada. 36 37 38 39 Study population 40 41 We identified a base cohort of all patients, at least 18 years of age, who were newly 42 43 44 prescribed an antihypertensive drug (including beta adrenoceptor blockers, alpha adrenoceptor 45 46 blockers, ACEIs, ARBs, calcium channel blockers (CCBs), vasodilators, centrally acting 47 48 antihypertensives, diuretics, ganglion blockers, and renin inhibitors) between January 1, 1988 49 50 and December 31, 2015. All patients were required to have at least one year of medical history in 51 52 53 the CPRD before their first ever antihypertensive drug prescription. 54 55 56 57 58 59 5 60 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 42
1 2 3 From the base cohort defined above, we identified a study cohort of all patients who 4 5 6 initiated a new antihypertensive drug class on or after January 1, 1995 (the first year where both 7 8 ACEIs and ARBs were available in the UK) until December 31, 2015. These patients included 9 10 those newly treated with an antihypertensive drug class (i.e., first ever antihypertensive 11 Confidential: For Review Only 12 prescriptions) as well as those who added on or switched to an antihypertensive drug class not 13 14 15 previously used in their treatment history. Cohort entry was defined as the date of this first 16 17 prescription. Patients with a previous diagnosis of any cancer (other than non melanoma skin 18 19 cancer) or those who previously received cancer treatments (chemotherapy or radiotherapy) at 20 21 22 any time before cohort entry were excluded. This was to ensure the identification of incident 23 24 cases of lung cancer during follow up, and to avoid the inclusion of patients with metastatic 25 26 lesions to the lung from other cancer sites. Finally, patients with less than one year follow up 27 28 29 after cohort entry were excluded for latency considerations and to ensure the identification of 30 31 incident events during follow up. 32 33 All patients meeting the study inclusion criteria were followed up starting one year after 34 35 cohort entry and until a diagnosis of incident lung cancer (identified on the basis of Read codes; 36 37 38 Supplementary Table 1) or censored upon death from any cause, end of registration with the 39 40 general practice, or the end of the study period (December 31, 2016), whichever occurred first. 41 42 43 44 45 Exposure Assessment 46 47 We used a time dependent exposure definition where each person day of follow up was 48 49 classified into one of three mutually exclusive exposure categories: ACEIs (alone or in 50 51 52 combination with other antihypertensive drugs, but no previous use of ARBs), ARBs (alone or in 53 54 combination with other non ACEI antihypertensive drugs), and other antihypertensive drugs 55 56 57 58 59 6 60 https://mc.manuscriptcentral.com/bmj Page 7 of 42 BMJ
1 2 3 (which also included patients who used both ACEIs and ARBs). A one year exposure lag period 4 5 6 was introduced to account for a minimum latency time window and to minimize reverse 7 8 causality. Thus, patients initiating an antihypertensive drug were considered unexposed until one 9 10 year after the date of the first prescription and considered exposed thereafter. The reference 11 Confidential: For Review Only 12 category were ARBs, as these drugs are recommended at the same disease stage, thereby 13 14 15 minimizing potential confounding by indication.[24] 16 17 18 19 Potential confounders 20 21 22 All models were adjusted for the following variables measured at cohort entry: age, sex, 23 24 year of cohort entry, BMI (<25 kg/m2, 25 29 kg/m2, ≥30 kg/m2, unknown), smoking status 25 26 (current, former, never, unknown), alcohol related disorders (including alcoholism, alcoholic 27 28 29 cirrhosis of the liver, alcoholic hepatitis and hepatic failure), and history of lung diseases 30 31 (including pneumonia, tuberculosis, chronic obstructive pulmonary disease [COPD]); all 32 33 measured at any time before cohort entry. In addition, models included duration of treated 34 35 hypertension (defined as the time between first ever prescription for an antihypertensive drug 36 37 38 and cohort entry) and use of statins at any time before cohort entry. Finally, the models were 39 40 adjusted for the total number of unique drug classes in the year before cohort entry, as a general 41 42 measure of comorbidity.[25] 43 44 45 46 47 Statistical analyses 48 49 Crude incidence rates of lung cancer and 95% confidence intervals (CIs) based on the 50 51 52 Poisson distribution were calculated for each exposure group. Time dependent Cox proportional 53 54 hazards models were used to estimate hazard ratios (HRs) and 95% CIs of lung cancer associated 55 56 57 58 59 7 60 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 42
1 2 3 with the use of ACEIs compared with the use of ARBs. The models were adjusted for the 4 5 6 potential confounders listed previously. 7 8 9 10 Secondary analyses 11 Confidential: For Review Only 12 We performed three secondary analyses. First, we assessed whether there was a duration 13 14 15 response relation between ACEI cumulative duration of use and lung cancer incidence. For this 16 17 time dependent analysis, HRs were estimated in a time dependent manner for three predefined 18 19 duration categories: ≤5 years, 5.1 10 years, and >10 years. Second, we also investigated the 20 21 22 association between time since initiation of ACEIs and lung cancer risk, estimating HRs for three 23 24 predefined categories: ≤5 years, 5.1 10 years, and >10 years. Finally, to investigate possible 25 26 effect modification by smoking status, we included an interaction term between the exposure and 27 28 29 smoking status variables. 30 31 32 33 Sensitivity analyses 34 35 Four sensitivity analyses were conducted to assess the robustness of our findings. First, 36 37 38 given uncertainties related to the length of the latency time window, we varied the length of the 39 40 exposure lag period to two and three years. Second, we repeated the primary analysis using 41 42 multiple imputation for variables with missing values (Supplementary Methods 1).[26,27] 43 44 45 Third, as an alternate means to control for confounding, we repeated the analysis by stratifying 46 47 the model on disease risk score deciles (Supplementary Methods 2). [28,29] Finally, we 48 49 repeated the analysis using a marginal structural Cox proportional hazards model using inverse 50 51 52 probability of treatment and censoring weighting; a method designed to adjust for time 53 54 55 56 57 58 59 8 60 https://mc.manuscriptcentral.com/bmj Page 9 of 42 BMJ
1 2 3 dependent confounding associated with time varying exposures (Supplementary Methods 4 5 6 3).[30,31] 7
8 9 10 Ancillary analyses 11 Confidential: For Review Only 12 We conducted two ancillary analyses to address the possibility that ARBs may be 13 14 15 associated with a decreased risk of lung cancer incidence.[32] The first compared ACEIs with 16 17 thiazide diuretics, as the latter have not been previously associated with lung cancer incidence. 18 19 Thus, for this analysis, exposure was redefined into four mutually exclusive categories: ACEIs 20 21 22 (alone or in combination with other antihypertensive drugs, but no previous use of thiazide 23 24 diuretics or ARBs), thiazide diuretics (alone or in combination with other non ACEI or non ARB 25 26 antihypertensive drugs), ARBs (alone or in combination with other non ACEI or non thiazide 27 28 29 antihypertensive drugs), and other antihypertensive drugs. The second analysis compared ARBs 30 31 to thiazide diuretics to assess whether the former are indeed associated with a decreased risk of 32 33 lung cancer. For both analyses, we assessed the association overall and by cumulative duration of 34 35 use. All analyses were conducted with SAS version 9.4 (SAS institute, Cary, NC). 36 37 38 39 40 Patient involvement 41 42 Our study was a secondary data analysis and did not include patients as study 43 44 45 participants. No patients were involved in setting the research question or the outcome measures, 46 47 nor were they involved in the design and implementation of the study. There are no plans to 48 49 involve patients in the dissemination of results, nor will we disseminate results directly to 50 51 52 patients, beyond our general media communications plan. 53 54 55 56 57 58 59 9 60 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 42
1 2 3 RESULTS 4 5 6 The cohort included 992,061 patients (Figure 1) followed for a mean (standard 7 8 deviation) of 6.4 (4.7) years beyond the one year post cohort entry latency period. During the 9 10 follow up period, 335,135 patients were prescribed ACEIs, 29,008 ARBs, and 101,637 both 11 Confidential: For Review Only 12 ACEIs and ARBs. The three most commonly used ACEIs were ramipril (26%), lisinopril (12%) 13 14 15 and perindopril (7%). Overall, 7952 patients were newly diagnosed with lung cancer during 16 17 6,350,584 person years of follow up, generating a crude incidence rate of 1.3 (95% CI: 1.2 to 18 19 1.3) per 1000 person years. 20 21 22 Table 1 presents baseline characteristics of the entire cohort, and by use of ACEIs and 23 24 ARBs at cohort entry. Compared with ARB users, ACEI users were more likely to be male, more 25 26 likely to have alcohol related disorders, to be current smokers, and to have a higher BMI. 27 28 29 Additionally, ACEI users had a shorter duration of treated hypertension and were more likely to 30 31 have used statins and other prescription drugs. Both ACEI and ARB users had a similar history 32 33 of pneumonia, tuberculosis, and COPD. 34 35 The results from primary and secondary analyses are presented in Table 2. Compared 36 37 38 with ARBs, ACEIs were associated with an overall 13% increased risk of lung cancer (1.6 vs 1.2 39 40 per 1000 person years; HR, 1.13; 95% CI, 1.00 to 1.29). In secondary analyses, the use of ACEIs 41 42 for less than 5 years was not associated with an increased risk of lung cancer (HR, 1.09; 95% CI, 43 44 45 0.96 to 1.24). However, the HR was elevated with 5 to 10 years of use (HR, 1.21; 95% CI, 1.05 46 47 to 1.39), and continued to increase with more than 10 years of use (HR, 1.31; 95% CI, 1.08 to 48 49 1.58). Similar associations were observed with time since ACEI initiation, with HRs increasing 50 51 52 with longer times since initiation, peaking at >10 years since initiation (HR, 1.28; 95% CI, 1.09 53 54 55 56 57 58 59 10 60 https://mc.manuscriptcentral.com/bmj Page 11 of 42 BMJ
1 2 3 to 1.51). Smoking status did not significantly modify the association between ACEI use and lung 4 5 6 cancer risk (p interaction=0.41; Supplementary Table 2). 7 8 9 10 Sensitivity analyses 11 Confidential: For Review Only 12 Results of sensitivity analyses are summarized in Figure 2 and Supplementary Tables 13 14 15 3-7. Overall, these yielded consistent results, generating HRs ranging between 1.12 and 1.26. 16 17 The latter estimate was from the marginal structural model that controlled for potential time 18 19 dependent confounding. 20 21 22 23 24 Ancillary analyses 25 26 Compared with the use of thiazide diuretics, the use of ACEIs was associated with a 6% 27 28 29 increased risk of lung cancer (HR, 1.06; 95% CI, 1.00 to 1.13) (Supplementary Table 8). 30 31 Similar to the main analysis, use of ACEIs for less than 5 years was not associated with an 32 33 increased risk of lung cancer, while HRs were elevated with increasing use, peaking with more 34 35 than 10 years of use (HR, 1.23; 95% CI, 1.05 to 1.45). Analysis comparing ARBs to thiazide 36 37 38 diuretics revealed null associations overall (HR, 0.98; 95% CI, 0.82 to 1.07) and by cumulative 39 40 duration of use (Supplementary Table9). 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 11 60 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 42
1 2 3 DISCUSSION 4 5 6 In this large population based study of nearly one million patients, the use of ACEIs was 7 8 associated with an overall 13% increased risk of lung cancer. Associations were evident after 5 9 10 years of use and increased with longer durations of use, particularly among those who used 11 Confidential: For Review Only 12 ACEIs for more than 10 years (31% increased risk). While the magnitudes of the observed 13 14 15 associations are modest, it is important to note that ACEIs are one of the most widely prescribed 16 17 drug classes; in the UK 70.1 million antihypertensives are dispensed each year, of which 18 19 approximately 32% are ACEIs.[33,34] Thus, small relative effects could translate into large 20 21 22 absolute numbers of patients at risk for lung cancer. Given the potential impact of our findings, it 23 24 is imperative that they are replicated in other settings, particularly among patients exposed for 25 26 longer durations. 27 28 29 30 31 Comparison with previous studies 32 33 The results of this study contrast with those of recent meta analyses of RCTs which 34 35 found no evidence of an increase of cancer.[8,9] However, these RCTs were not designed to 36 37 38 assess cancer as outcome (and specifically lung cancer) and where the median duration of 39 40 follow up was 3.5 years (range 1.3 to 5.1 years).[8,9] Indeed, in this study, associations with 41 42 lung cancer risk were only evident with over 5 years of ACEI use. To our knowledge, while 43 44 45 several observational studies reported on the association between ACEIs and lung cancer 46 47 incidence,[10–16] only one study was specifically designed to investigate this association.[17] In 48 49 this well conducted study, the use of ACEIs was not associated with an increased risk of lung 50 51 52 cancer (HR, 0.99; 95% CI, 0.84 to 1.16), when compared with ARBs. However, as this study had 53 54 a maximum follow up of 5 years, its conclusion is not incompatible with our finding suggesting 55 56 57 58 59 12 60 https://mc.manuscriptcentral.com/bmj Page 13 of 42 BMJ
1 2 3 no association in the first 5 years of use (HR, 1.09; 95% CI, 0.96 to 1.24). While other 4 5 6 observational studies investigated this association, their findings were part of secondary analyses 7 8 and thus should be interpreted with caution. Overall, these studies produced mixed results, with 9 10 some reporting increased risks,[10,11,16] others null associations,[12–14] and one study 11 Confidential: For Review Only 12 reporting a 66% decreased risk.[15] However, the latter study may have suffered from immortal 13 14 15 time bias, which resulted from the misclassification of unexposed person time as exposed 16 17 person time.[35] The other studies had other limitations, such as the inclusion of prevalent users 18 19 of antihypertensive drugs,[15] confounding by indication,[14] and not accounting for cancer 20 21 22 latency in their analyses.[13,15,17] 23 24 The association between ACEIs and lung cancer is biologically plausible. In addition to 25 26 angiotensin I, ACE also metabolizes bradykinin, an active vasodilator.[36] Thus, the use of 27 28 29 ACEIs results in the accumulation of bradykinin in the lung.[5] Bradykinin receptors have been 30 31 located on various cancer tissues including lung cancer,[5,37] and it has been reported that 32 33 bradykinin may directly stimulate growth of lung cancer.[5,6] Bradykinin has been shown to 34 35 stimulate the release of vascular endothelial growth factor (VEGF), thus promoting 36 37 38 angiogenesis,[38,39] as well as indirect effects on lung cancer by enhancing vascular 39 40 permeability, via the activation of MMP, facilitating tumor invasion and metastases.[39] 41 42 Moreover, ACEI use also results in accumulation of substance P which is expressed in lung 43 44 45 cancer tissue and is associated with tumor proliferation and angiogenesis.[7] 46 47 The results of this study also raise important questions regarding the new angiotensin 48 49 receptor/neprilysin inhibitor sacubitril/valsartan. Neprilysin inhibition results in increases in 50 51 52 vasoactive and other peptides including bradykinin and substance P.[40] The recent 53 54 PARADIGM trial reported clinical benefits for cardiovascular outcomes and death, however 55 56 57 58 59 13 60 https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 42
1 2 3 cancer events were not reported.[41] Therefore, it remains unknown whether these new renin 4 5 6 angiotensin system (RAS) inhibitors may also increase the risk of lung cancer in the long term. 7 8 9 10 Strengths and Limitations 11 Confidential: For Review Only 12 This study has several strengths. First, to our knowledge, with over 990,000 patients 13 14 15 followed for an average of 6.4 tears (beyond the one year post cohort entry lag period), this 16 17 study is the largest to have been conducted to specifically assess this association. Second, we 18 19 used a new user design, thus minimizing biases related to the inclusion of prevalent users.38 20 21 22 Third, we used a time dependent exposure definition that eliminated immortal time bias while 23 24 also accounting for cancer latency. Finally, the use of the CPRD allowed us to adjust the models 25 26 for a number of potential important confounders, including smoking status, which was not 27 28 29 available in some of the previous studies.[17] 30 31 This study has some limitations. First, prescriptions in the CPRD represent those written 32 33 by general practitioners, and thus misclassification of exposure is possible if patients did not 34 35 comply with the treatment regimen or received prescriptions from specialists. However, since all 36 37 38 patients entering the cohort were those newly treated with antihypertensive drugs, 39 40 misclassification due to non adherence should be minimal and likely non differential between 41 42 ACEI and ARB users. Second, we compared ACEIs to ARBs, since the latter also act upon the 43 44 45 RAS and are used at the same disease stage, but have not been associated with neuropeptide 46 47 accumulation in the lung. However, it has been suggested that ARBs may also have an effect on 48 49 lung cancer incidence,[43] including a meta analysis of observational studies reporting a 50 51 52 decreased risk with these drugs.[32] Studies included in this meta analysis had a number of 53 54 limitations, and several compared ARBs to ACEIs.[32] Thus, the apparent protective effect of 55 56 57 58 59 14 60 https://mc.manuscriptcentral.com/bmj Page 15 of 42 BMJ
1 2 3 ARBs may be the result of a deleterious effect of ACEIs on lung cancer incidence.[32] 4 5 6 Nonetheless, our study was designed to address this possibility by comparing ACEIs to thiazide 7 8 diuretics in ancillary analyses. Reassuringly, this analysis yielded consistent results, both in 9 10 terms of overall association and by cumulative duration of use. Importantly, our analyses 11 Confidential: For Review Only 12 comparing ARBs to thiazide diuretics produced null associations for both overall and cumulative 13 14 15 duration of use; this suggests that the observed increased risk with ACEIs is unlikely to be 16 17 attributed to the purported antitumor effects of ARBs. Third, misclassification of the outcome is 18 19 possible; however, lung cancer has been shown to be well recorded in the CPRD when compared 20 21 22 with the UK National Cancer Data Repository (concordance rate of 93%).[23] Associations may 23 24 also vary by lung cancer subtypes, but this information was not available within CPRD. Finally, 25 26 persistent cough is a common and well known side effect of ACEIs, raising the possibility that 27 28 29 the observed association could be due to detection bias. Indeed, it is possible that patients on 30 31 ACEIs may be more likely to undergo diagnostic evaluations, such as computerized tomography 32 33 of chest, leading to an increased detection of preclinical lung cancers. Information on chest 34 35 workup is not well recorded in CPRD, and therefore it was not possible to account for this 36 37 38 possibility in our analyses. However, a recent study found minimal evidence of differences in 39 40 chest workup after ACEI and ARB initiation.[17] Moreover, an over detection of lung cancer 41 42 would be expected to be observed relatively soon after treatment initiation, which is one the 43 44 45 reasons why our exposures were lagged by one year. Lengthening the exposure lag period to two 46 47 and three year yielded consistent findings as those observed for the primary analysis. 48 49 Furthermore, associations between ACEI use and lung cancer risk were evident only with 50 51 52 increasing durations of use (>5 years). Taken together, these results do not corroborate the 53 54 hypothesis of an over detection of lung cancer among ACEIs 55 56 57 58 59 15 60 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 42
1 2 3 Conclusions 4 5 6 In this large, population based study, the use of ACEIs was associated with an elevated 7 8 risk of lung cancer overall, along with evidence of a duration response relation. While these 9 10 findings need to be replicated in other settings, physicians need to be aware of this possible 11 Confidential: For Review Only 12 association. 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 16 60 https://mc.manuscriptcentral.com/bmj Page 17 of 42 BMJ
1 2 3 What is already known on this subject 4 5 6 • Biological evidence suggests that angiotensin converting enzyme inhibitors may increase 7 the risk of lung cancer via the accumulation of bradykinin and substance P in the lung. 8 9 10 11 • HoweverConfidential: observational studies examining For this Review association are limited Only and report 12 13 inconsistent results. 14 15 16 What this study adds 17 18 • The use of angiotensin converting enzyme inhibitors was associated with a 13% 19 20 increased risk of lung cancer. 21 22 23 24 • Associations were evident after 5 years of use and increased with longer durations of use, 25 particularly among those who used ACEIs for more than 10 years. 26 27 28 29 30 31 • While the magnitudes of the observed estimates are modest, these small relative effects 32 could translate into large absolute numbers of patients at risk for lung cancer, thus it is 33 34 imperative that these findings are replicated in other settings. 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 17 60 https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 42
1 2 3 ACKNOWLEDGMENTS 4 5 6 Author Contributions 7 8 All authors conceived and designed the study, analysed and interpreted the data, and critically 9 10 revised the manuscript for important intellectual content. LA acquired the data. BH HY and LA 11 Confidential: For Review Only 12 did the analyses. BH wrote the manuscript and all authors participated in the interpretation of the 13 14 15 results and critical revision of the manuscript. LA is the guarantor. 16 17 18 19 Funding/Support 20 21 22 This study was funded by a Foundation Scheme grant from the Canadian Institutes of Health 23 24 Research. Dr Blánaid Hicks holds a Cancer Research UK Population Research Fellowship. Dr. 25 26 Kristian Filion holds a Chercheur Boursier Junior 2 award from the Fonds de recherche du 27 28 29 Québec – Santé (FRQS). Dr. Jay Udell holds a Heart and Stroke Foundation of Canada National 30 31 New Investigator/Ontario Clinician Scientist (Phase I) Award. Dr. Laurent Azoulay holds a 32 33 Chercheur Boursier Junior 2 award from the FRQS and is the recipient of a William Dawson 34 35 Scholar award from McGill University. The funding source had no influence on the design and 36 37 38 conduct of the study; collection, management, analysis, and interpretation of the data; 39 40 preparation, review, or approval of the manuscript; and decision to submit the manuscript for 41 42 publication. 43 44 45 46 47 Competing Interests 48 49 All authors have completed the ICMJE uniform disclosure form at 50 51 52 www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and 53 54 declare: this study was funded by the Canadian Institutes of Health Research; no financial 55 56 57 58 59 18 60 https://mc.manuscriptcentral.com/bmj Page 19 of 42 BMJ
1 2 3 relationships with any organisations that might have an interest in the submitted work in the 4 5 6 previous three years; no other relationships or activities that could appear to have influenced the 7 8 submitted work. 9 10 11 Confidential: For Review Only 12 Details of Ethical Approval 13 14 15 The study protocol was approved by the Independent Scientific Advisory Committee of the 16 17 Clinical Practice Research Datalink (protocol number 16_255R) and by the Research Ethics 18 19 Board of Jewish General Hospital, Montreal, Quebec, Canada. 20 21 22 Data Sharing 23 24 No additional data available. 25 26 27 Transparency 28 29 The guarantor (LA) affirms that this manuscript is an honest, accurate, and transparent account 30 31 of the study being reported; that no important aspects of the study have been omitted; and that 32 33 any discrepancies from the study as planned (and, if relevant, registered) have been explain 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 19 60 https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 42
1 2 3 REFERENCES 4 5 1 Heran BS, Wong MM, Heran IK, et al. Blood pressure lowering efficacy of angiotensin 6 7 converting enzyme (ACE) inhibitors for primary hypertension. Cochrane database Syst 8 9 10 Rev 2008;:CD003823. doi:10.1002/14651858.CD003823.pub2 11 Confidential: For Review Only 12 2 Chan Yoon MD, Hyun Sik Yang MD, Inpyo Jeon MD, Yoosoo Chang MD SMPMP. Use 13 14 of angiotensin converting enzyme inhibitors or angiotensin receptor blockers and cancer 15 16 17 risk: a meta analysis of observational studies. CMAJ Published Online First: 2011. 18 19 doi:DOI:10.1503 /cmaj.101497 20 21 3 Rosenthal T, Gavras I. Angiotensin inhibition and malignancies: a review. J Hum 22 23 24 Hypertens 2009;23:623–35. doi:10.1038/jhh.2009.21 25 26 4 Shen J, Huang Y M, Wang M, et al. Renin angiotensin system blockade for the risk of 27 28 cancer and death. J Renin-Angiotensin-Aldosterone Syst 2016;17. 29 30 doi:10.1177/1470320316656679 31 32 33 5 Trifilieff A, Da Silva A, Gies JP. Kinins and respiratory tract diseases. Eur Respir J 34 35 1993;6:576–87. 36 37 6 Sethi T, Rozengurt E. Multiple neuropeptides stimulate clonal growth of small cell lung 38 39 40 cancer: effects of bradykinin, vasopressin, cholecystokinin, galanin, and neurotensin. 41 42 Cancer Res 1991;51:3621–3. 43 44 7 Muñoz M, Coveñas R. Involvement of substance P and the NK 1 receptor in human 45 46 47 pathology. Amino Acids 2014;46:1727–50. doi:10.1007/s00726 014 1736 9 48 49 8 ARB Trialists Collaboration. Effects of telmisartan, irbesartan, valsartan, candesartan, and 50 51 losartan on cancers in 15 trials enrolling 138,769 individuals. J Hypertens 2011;29:623– 52 53 35. doi:10.1097/HJH.0b013e328344a7de 54 55 56 9 Bangalore S, Kumar S, Kjeldsen SE, et al. Antihypertensive drugs and risk of cancer: 57 58 59 20 60 https://mc.manuscriptcentral.com/bmj Page 21 of 42 BMJ
1 2 3 network meta analyses and trial sequential analyses of 324 168 participants from 4 5 6 randomised trials. Lancet Oncol 2011;12:65–82. doi:10.1016/S1470 2045(10)70260 6 7 8 10 Azoulay L, Assimes TL, Yin H, et al. Long term use of angiotensin receptor blockers and 9 10 the risk of cancer. PLoS One 2012;7:e50893. doi:10.1371/journal.pone.0050893 11 Confidential: For Review Only 12 11 Bhaskaran K, Douglas I, Evans S, et al. Angiotensin receptor blockers and risk of cancer: 13 14 15 cohort study among people receiving antihypertensive drugs in UK General Practice 16 17 Research Database. BMJ 2012;344:e2697. 18 19 12 Chang C H, Lin J W, Wu L C, et al. Angiotensin receptor blockade and risk of cancer in 20 21 22 type 2 diabetes mellitus: a nationwide case control study. J Clin Oncol 2011;29:3001–7. 23 24 doi:10.1200/JCO.2011.35.1908 25 26 13 Friis S, Sorensen HT, Mellemkjaar L, et al. Angiotensin converting enzyme inhibitors and 27 28 29 the risk of cancer. Cancer 2001;92:2462–70. 30 31 14 Hallas J, Christensen R, Andersen M, et al. Long term use of drugs affecting the renin 32 33 angiotensin system and the risk of cancer: a population based case control study. Br J Clin 34 35 Pharmacol 2012;74:180–8. doi:10.1111/j.1365 2125.2012.04170.x 36 37 38 15 Lever AF, Hole DJ, Gillis CR, et al. Do inhibitors of angiotensin I converting enzyme 39 40 protect against risk of cancer? Lancet 1998;352:179–84. doi:10.1016/S0140 41 42 6736(98)03228 0 43 44 45 16 Pasternak B, Svanström H, Callréus T, et al. Use of angiotensin receptor blockers and the 46 47 risk of cancer. Circulation 2011;123:1729–36. 48 49 17 Gokhale M, Girman C, Chen Y, et al. Comparison of diagnostic evaluations for cough 50 51 52 among initiators of angiotensin converting enzyme inhibitors and angiotensin receptor 53 54 blockers. Pharmacoepidemiol Drug Saf 2016;25:512–20. doi:10.1002/pds.3977 55 56 57 58 59 21 60 https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 42
1 2 3 18 Herrett E, Gallagher AM, Bhaskaran K, et al. Data Resource Profile: Clinical Practice 4 5 6 Research Datalink (CPRD). Int J Epidemiol 2015;44:827–36. doi:10.1093/ije/dyv098 7 8 19 National Health Service. Read Codes. http://systems.digital.nhs.uk/data/uktc/readcodes 9 10 [Accessed October 19 2017] 11 Confidential: For Review Only 12 20 NHS Business Services Authority. Dictionary of medicines and devicess. 13 14 15 https://www.nhsbsa.nhs.uk/pharmacies gp practices and appliance contractors/dictionary 16 17 medicines and devices dmd [Accessed October 19, 2017]. 18 19 21 Jick SS, Kaye JA, Vasilakis Scaramozza C, et al. Validity of the general practice research 20 21 22 database. Pharmacotherapy 2003;23:686–9. 23 24 22 Lawrenson R, Williams T, Farmer R. Clinical information for research; the use of general 25 26 practice databases. J Public Health Med 1999;21:299–304. 27 28 29 23 Boggon R, van Staa TP, Chapman M, et al. Cancer recording and mortality in the General 30 31 Practice Research Database and linked cancer registries. Pharmacoepidemiol Drug Saf 32 33 2013;22:168–75. doi:10.1002/pds.3374 34 35 24 The National Institute for Health and Care Excellence. Hypertension in adults: diagnosis 36 37 38 and management. https://www.nice.org.uk/guidance/cg127/chapter/1 Guidance#initiating 39 40 and monitoring antihypertensive drug treatment including blood pressure targets 2. 41 42 [Accessed October 23, 2017]. 43 44 45 25 Schneeweiss S, Seeger JD, Maclure M, et al. Performance of comorbidity scores to 46 47 control for confounding in epidemiologic studies using claims data. Am J Epidemiol 48 49 2001;154:854–64. 50 51 52 26 DB R. Multiple Imputation for Nonresponse in Surveys. Hoboken, NJ, USA: : John Wiley 53 54 & Sons, Inc. 1987. doi:10.1002/9780470316696 55 56 57 58 59 22 60 https://mc.manuscriptcentral.com/bmj Page 23 of 42 BMJ
1 2 3 27 Schafer JL. Analysis of Incomplete Multivariate Data. Chapman and Hall 1997. 4 5 6 28 Arbogast PG, Ray WA. Performance of disease risk scores, propensity scores, and 7 8 traditional multivariable outcome regression in the presence of multiple confounders. Am 9 10 J Epidemiol 2011;174:613–20. doi:10.1093/aje/kwr143 11 Confidential: For Review Only 12 29 Arbogast PG, Ray WA. Use of disease risk scores in pharmacoepidemiologic studies. Stat 13 14 15 Methods Med Res 2009;18:67–80. doi:10.1177/0962280208092347 16 17 30 Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate the causal 18 19 effect of zidovudine on the survival of HIV positive men. Epidemiology 2000;11:561–70. 20 21 22 31 Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in 23 24 epidemiology. Epidemiology 2000;11:550–60. 25 26 32 Zhang J, Liu J, Chen J, et al. Angiotensin receptor blockers (ARBs) reduce the risk of 27 28 29 lung cancer: a systematic review and meta analysis. Int J Clin Exp Med 2015;8:12656–60. 30 31 33 Fretheim A, Oxman AD. International variation in prescribing antihypertensive drugs: Its 32 33 extent and possible explanations. BMC Health Serv Res doi:10.1186/1472 6963 5 21 34 35 34 NHS Digital. Prescriptions Dispensed in the Community, Statistics for England 2004 36 37 38 2014. https://digital.nhs.uk/catalogue/PUB17644 [September 28 2017]. 39 40 35 Suissa S. Immortal time bias in pharmaco epidemiology. Am J Epidemiol 2008;167:492– 41 42 9. doi:10.1093/aje/kwm324 43 44 45 36 Carey RM, Siragy HM. Newly recognized components of the renin angiotensin system: 46 47 potential roles in cardiovascular and renal regulation. Endocr Rev 2003;24:261–71. 48 49 doi:10.1210/er.2003 0001 50 51 52 37 Golias C, Charalabopoulos A, Stagikas D, et al. The kinin system bradykinin: biological 53 54 effects and clinical implications. Multiple role of the kinin system bradykinin. 55 56 57 58 59 23 60 https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 42
1 2 3 Hippokratia 2007;11:124–8. 4 5 6 38 Ishihara K, Hayash I, Yamashina S, et al. A potential role of bradykinin in angiogenesis 7 8 and growth of S 180 mouse tumors. Jpn J Pharmacol 2001;87:318–26. 9 10 39 Stewart JM. Bradykinin antagonists as anti cancer agents. Curr Pharm Des 2003;9:2036– 11 Confidential: For Review Only 12 42. 13 14 15 40 Singh JSS, Burrell LM, Cherif M, et al. Sacubitril/valsartan: beyond natriuretic peptides. 16 17 Heart 2017;103:1569–77. doi:10.1136/heartjnl 2017 311295 18 19 41 McMurray JJV, Packer M, Desai AS, et al. Angiotensin–Neprilysin Inhibition versus 20 21 22 Enalapril in Heart Failure. N Engl J Med 2014;371:993–1004. 23 24 doi:10.1056/NEJMoa1409077 25 26 42 Ray WA. Evaluating medication effects outside of clinical trials: new user designs. Am J 27 28 29 Epidemiol 2003;158:915–20. 30 31 43 Sipahi I, Debanne SM, Rowland DY, et al. Angiotensin receptor blockade and risk of 32 33 cancer: meta analysis of randomised controlled trials. doi:10.1016/S1470 2045(10)70106 34 35 6 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 24 60 https://mc.manuscriptcentral.com/bmj Page 25 of 42 BMJ
1 2 3 FIGURE LEGENDS 4 5 6 7 8 Figure 1 Study flow diagram describing the construction of the base and study cohorts 9 10 ARB indicates Angiotensin receptor blockers 11 Confidential: For Review Only 12 13 14 15 Figure 2 Forrest plot summarizing the results of the primary and sensitivity analyses 16 17 assessing the association between ACEI use and lung cancer incidence. 18 19 HR indicates hazard ratio; CI , confidence interval 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 25 60 https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 42
1 Figure 1 2 3 4 5 2,989,568 Patients with a first ever
6 prescription for an 7 antihypertensive drug 8 between January 1, 1988 to 9 December 31, 2015 10 Confidential: For Review Only 11 12 13 1,776,955 Excluded 14 15 36,229 < 18 years of age 16 1,736,618 < 365 days coverage in the database 17 4108 Date inconsistencies 18 19 20 21 22 23 1,212,613 Patients included in the base 24 cohort 25 26 27 44,527 Excluded 28 29 11,440 Died or left cohort before first ARB entered the market 30 33,087 Never added on or switched to new antihypertensive drug 31 class after ARB entered the market 32 33 34
35 1,168,086 Cohort of new users or 36 switchers after ARBs 37 entered the market 38 39 40 41 42 43 176,025 Excluded 44 80,688 Previous diagnosis of cancer or receipt of cancer 45 treatments 46 95,337 Less than 1 year of follow up 47 48 49 50 51
52 992,061 Study cohort 53 54 55 56 57 58 59 26 60 https://mc.manuscriptcentral.com/bmj Page 27 of 42 BMJ
1 2 3 Table 1. Baseline Demographic and Clinical Characteristics of the Cohort and Stratified by drug use at 4 cohort entry 5 6 Use at Cohort Entry 7 Characteristic Entire Cohort ACEIs ARBs 8 Total 992,061 208,353 (21.0) 16,027 (1.6) 9 Age, years (mean, SD) 55.6 (16.6) 57.8 (13.1) 57.9 (13.2) 10 Male, n (%) Confidential: 459,064For (46.3) Review 133,091 (63.9) Only 9591 (59.8) 11 Alcohol related disorders, n (%) 71,605 (7.2) 18,199 (8.7) 1092 (6.8) 12 13 Smoking status, n (%) 14 Current 215,098 (21.7) 41,595 (20.0) 2802 (17.5) 15 Past 227,504 (22.9) 58,683 (28.2) 3916 (24.4) 16 Never 484,831 (48.9) 99,820 (47.9) 8248 (51.5) 17 Unknown 64,628 (6.5) 8255 (4.0) 1061 (6.6) 18 Body mass index, n (%) 19 < 25 kg/m2 303,311 (30.6) 45,164 (21.7) 3602 (22.5) 20 2 21 25 30 kg/m 304,699 (30.7) 71,655 (34.4) 5447 (34.0) 22 ≥30.0 224,888 (22.7) 67,353 (32.3) 4724 (29.5) 23 Unknown 159,163 (16.0) 24,181 (11.6) 2254 (14.1) 24 Duration of treated hypertension (mean, SD) 0.2 (1.5) 0.3 (1.8) 0.5 (2.4) 25 Pneumonia, n (%) 22,403 (2.3) 5027 (2.4) 320 (2.0) 26 Tuberculosis, n (%) 2399 (0.2) 474 (0.2) 37 (0.2) 27 28 Chronic obstructive pulmonary disease, n (%) 78,669 (7.9) 16,152 (7.8) 1180 (7.4) 29 Statins, n (%) 164,891 (16.6) 73,510 (35.3) 4092 (25.5) 30 Total number of unique drug classes, n (%) 4.1 (4.1) 4.1 (4.1) 3.8 (4.1) 31 0 150,293 (15.2) 35,384 (17.0) 3107 (19.4) 32 1 147,609 (14.9) 31,022 (14.9) 2603 (16.2) 33 2 135,085 (13.6) 27,027 (13.0) 2195 (13.7) 34 3 115,121 (11.6) 22,157 (10.6) 1740 (10.9) 35 36 ≥4 443,953 (44.7) 92,763 (44.5) 6382 (39.8) 37 Abbreviations: SD, standard deviation ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers. 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 27 60 https://mc.manuscriptcentral.com/bmj Page 28 of 42 28 28 ‡ ‡ for trend= 0.0002 trend= for for trend= 0.0002 trend= for P
Adjusted HR HR Adjusted (95% CI) P
Crude HR Crude
† †
(95% CI) Incidence rate rate Incidence
BMJ
years
Person- https://mc.manuscriptcentral.com/bmj 414 414 171,596 to 2.7) (2.2 2.4 1.62 to 1.51) (1.09 1.28
1617 1617 1,158,441 to 1.5) (1.3 1.4 1.24 to 1.26) (0.97 1.10
1155 1155 647,103 to 1.9) (1.7 1.8 1.33 to1.29) (0.99 1.13 2084 2084 1,440,232 to 1.5) (1.4 1.4 1.24 to 1.24) (0.96 1.09 197 197 79,598 to 2.8) (2.1 2.5 1.63 to 1.58) (1.08 1.31
266 266 3186 213,557 1,977,139 to 1.4) (1.1 1.2 to 1.7) (1.6 1.6 1.00 1.32 to 1.29) (1.00 1.13 [Reference] 1.00 905 905 457,309 to 2.1) (1.9 2.0 1.44 to 1.39) (1.05 1.21 Events Events
Confidential: For Review Only
* Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders, history of lung diseases prior to cohort entry entry cohort to prior diseases of lung history disorders, alcohol related smoking, index, mass body entry, cohort of sex, year age, for Adjusted Per 1000 Person Years. Person Years. 1000 Per Use of other antihypertensive drugs is considered in the model, but not presented in the table. table. in the presented not but model, the in considered is drugs antihypertensive of other Use (including pneumonia, tuberculosis, and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the and statins, of use hypertension, treated of duration disease), pulmonary obstructive chronic of history and tuberculosis, pneumonia, (including entry. cohort before year the classes in drug unique of total number Abbreviations: HR, hazard ratio; CI, confidence interval; ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers blockers receptor Angiotensin ARB, inhibitors; enzyme converting Angiotensin ACEI, interval; confidence CI, ratio; hazard HR, Abbreviations: * † ‡ ≤ 5 ≤ 5 5.1 10 ≤ 5 ≤ 5 5.1 10 > 10 10 > > 10 10 >
Table 2. Crude and Adjusted HRs for the Association Between the Use of ACEIs and the Risk of Lung Cancer of Cancer Lung Risk of and the ACEIs Use the Between Association for the HRs and Adjusted Crude Table 2. Exposure use ACEI first (years) since Time
Cumulative duration of ACEI use (years) (years) of use duration ACEI Cumulative
ARBs ACEIs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 42 BMJ
1 2 3 Figure 2 4 5 6 7 8 9 10 11 Confidential: For Review Only 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 29 60 https://mc.manuscriptcentral.com/bmj BMJ Page 30 of 42
1 2 3 Supplementary Online Content 4 5 Supplementary Method 1 Multiple Imputation 6 7 Supplementary Method 2 Disease Risk Score Analysis 8 9 Supplementary Method 3 Marginal Structural Cox Proportional Hazards Model 10 Confidential: For Review Only 11 Supplementary Table 1 Read Codes for Lung Cancer Diagnosis 12 13 14 Supplementary Table 2 Crude and Adjusted HRs for the Association Between the Use of 15 ACEI and the Risk of Lung Cancer Interaction with Smoking) 16 17 Supplementary Table 3 Crude and Adjusted HRs for the Association Between the Use of 18 ACEI and the Risk of Lung Cancer (Varying the Lag Period to 2 19 Years) 20 21 Supplementary Table 4 Crude and Adjusted HRs for the Association Between the Use of 22 ACEI and the Risk of Lung Cancer (Varying the Lag Period to 3 23 Years) 24 25 Supplementary Table 5 Crude and Adjusted HRs for the Association Between the Use of 26 ACEI and the Risk of Lung Cancer (Multiple Imputation) 27 28 29 Supplementary Table 6 Crude and Adjusted HRs for the Association Between the Use of 30 ACEI and the Risk of Lung Cancer (Disease Risk Score Analysis) 31 32 Supplementary Table 7 Crude and Adjusted HRs for the Association Between the Use of 33 ACEI and the Risk of Lung Cancer (Marginal Structural Model) 34 35 Supplementary Table 8 Crude and Adjusted HRs for the Association Between the Use of 36 ACEI and the Risk of Lung Cancer (Using Thiazide Diuretic Use as 37 an Alternative Comparator Group) 38 39 Supplementary Table 9 Crude and Adjusted HRs for the Association Between the Use of 40 41 ARBs Compared to Thiazide Diuretic Use and the Risk of Lung. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 1 60 https://mc.manuscriptcentral.com/bmj Page 31 of 42 BMJ
1 2 3 Supplementary Method 1 Multiple Imputation 4 5 The primary analysis was repeated using multiple imputation for variables with 1,2 6 missing values. An ordinal logistic regression model was used to impute variables 7 with missing information (including smoking and BMI) with explanatory variables 8 and cumulative hazard (as recommended 3 and ACEI use at cohort entry), along with 9 all confounders mentioned previously. Ten imputations were conducted, and the 10 results combined using Rubin’s rules.4 11 Confidential: For Review Only 12 References 13 14 1 Schafer JL. Analysis of Incomplete Multivariate Data. Chapman and Hall 1997. 15 16 2 Rubin DB, editor. Multiple Imputation for Nonresponse in Surveys. Hoboken, NJ, 17 USA: : John Wiley & Sons, Inc. 1987. doi:10.1002/9780470316696 18 19 3 White IR, Royston P. Imputing missing covariate values for the Cox model. Stat 20 Med 2009;28:1982–98. doi:10.1002/sim.3618 4 Sterne JAC, White IR, Carlin JB, et 21 al. Multiple imputation for missing data in epidemiological and clinical research: 22 potential and pitfalls. BMJ 2009;338:b2393. 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 2 60 https://mc.manuscriptcentral.com/bmj BMJ Page 32 of 42
1 2 3 Supplementary Method 2. Disease Risk Score Analysis 4 5 Disease Risk Score Analysis (DRS) was conducted, utilizing a historical cohort to 1 6 address residual confounding. This is analogous to the propensity score, and has been 7 shown to have a comparable performance.2,3 Using the same study inclusion criteria 8 for the primary analyses, a cohort of patients newly prescribed antihypertensive 9 medications from January 1 1990 and December 31st 1994 was first identified. A Cox 10 proportional hazards model was fitted that included all potential baseline confounders 11 Confidential: For Review Only 12 listed in the manuscript (excluding statin use and tuberculosis diagnosis due to a low 13 number of events). The disease risk score was then calculated in the historical cohort 14 by taking the products of the regression coefficients and the individual covariate 15 values (but where the exposure status was set to zero), and summing these products 16 for each patient. Hazard ratios of lung cancer were estimated by comparing use of 17 ACEIs with use of ARBs by stratifying the Cox proportional hazards model on DRS 18 19 deciles. 20 21 22 References 23 24 1. Glynn, Gagne & Schneeweiss. Role of disease risk score in comparative 25 effectiveness research with emerging therapies. Pharmacoepidemiol Drug Saf 26 2012; 21(s2):138 147 27
28 29 2. Arbogast PG, Ray WA. Performance of disease risk scores, propensity scores, 30 and traditional multivariable outcome regression in the presence of multiple 31 confounders. Am J Epidemiol. 2011;174(5):613 620. doi:10.1093/aje/kwr143. 32 33 3. Arbogast PG, Ray WA. Use of disease risk scores in pharmacoepidemiologic 34 studies. Stat Methods Med Res. 2009;18(1):67 80. 35 36 Doi:10.1177/0962280208092347. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 3 60 https://mc.manuscriptcentral.com/bmj Page 33 of 42 BMJ
1 2 3 Supplementary Method 3. Marginal Structural Cox Proportional Hazards 4 Model. 5 6 We conducted a marginal structural Cox proportional hazards model to address 7 potential residual time dependent confounding over the 21 year follow up period. 8 This method is designed to adjust for time dependent confounding associated with 9 time varying exposures.1,2 Two pooled logistic regression models were fitted to 10 estimate the conditional probability of being exposed to ACEIs at 30 day intervals 11 Confidential: For Review Only 12 during follow up; one for the numerator and one for the denominator of the stabilized 13 inverse probability of treatment weights (IPTWs). The numerator treatment model 14 included baseline covariates (the covariates included in the primary analysis, listed in 15 the main text) and follow up time. The denominator model included covariates (the 16 same as those listed in the main text) measured at each time interval and follow up 17 time. In both the numerator and denominator models, follow up time was modelled 18 19 using a restricted cubic spline with five knots to reduce bias due to model 3 20 misspecification from linearity assumptions. We also estimated inverse probability of 21 censoring weights (IPCWs) in a similar fashion. Stabilized IPTW and IPCW for each 22 patient in the cohort were computed using the predictive probabilities from both 23 treatment and censoring models. The product of these stabilized IPTWs and IPCWs 24 was then used to reweight the cohort, in which we estimated the marginal HR of lung 25 26 cancer associated with ACEI use, with 95% confidence intervals, calculated using 2 27 robust variance estimators. 28 29 30 References 31 32 1. Robins JM, Hernán MA, Brumback B. Marginal structural models and causal 33 inference in epidemiology. Epidemiology 2000;11:550–60. 34
35 36 2. Hernán MA, Brumback B, Robins JM. Marginal structural models to estimate 37 the causal effect of zidovudine on the survival of HIV positive men. 38 Epidemiology 2000;11:561–70. 39 40 3. Cole SR, Hernán MA. Constructing inverse probability weights for marginal 41 42 structural models. Am J Epidemiol 2008;168:656–64. doi:10.1093/aje/kwn164 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 4 60 https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 42
1 2 3 Supplementary Table 1 Read Codes for Lung Cancer Diagnosis 4 5 Read Code Read Term 6 B22z.11 Lung cancer 7 B22z.00 Malignant neoplasm of bronchus or lung NOS 8 B907200 Neoplasm of uncertain behaviour of lung 9 H58y400 Squamous metaplasia of lung 10 BB1K.00 [M]Oat cell carcinoma 11 Confidential: For Review Only 12 B222.00 Malignant neoplasm of upper lobe, bronchus or lung 13 B224100 Malignant neoplasm of lower lobe of lung 14 B221.00 Malignant neoplasm of main bronchus 15 B22..00 Malignant neoplasm of trachea, bronchus and lung 16 B221000 Malignant neoplasm of carina of bronchus 17 18 B224000 Malignant neoplasm of lower lobe bronchus 19 B222.11 Pancoast's syndrome 20 B221z00 Malignant neoplasm of main bronchus NOS 21 B222100 Malignant neoplasm of upper lobe of lung 22 B224.00 Malignant neoplasm of lower lobe, bronchus or lung 23 B223.00 Malignant neoplasm of middle lobe, bronchus or lung 24 25 B222000 Malignant neoplasm of upper lobe bronchus 26 B221100 Malignant neoplasm of hilus of lung 27 B225.00 Malignant neoplasm of overlapping lesion of bronchus & lung 28 B22y.00 Malignant neoplasm of other sites of bronchus or lung 29 B223100 Malignant neoplasm of middle lobe of lung 30 Byu2000 [X]Malignant neoplasm of bronchus or lung, unspecified 31 32 B223000 Malignant neoplasm of middle lobe bronchus 33 B224z00 Malignant neoplasm of lower lobe, bronchus or lung NOS 34 B222z00 Malignant neoplasm of upper lobe, bronchus or lung NOS 35 B223z00 Malignant neoplasm of middle lobe, bronchus or lung NOS 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 5 60 https://mc.manuscriptcentral.com/bmj 6
:0.41 interaction - p * 1.00 (reference) (reference) 1.00 (reference) 1.00
BMJ
https://mc.manuscriptcentral.com/bmj (Interaction with Smoking) with (Interaction
Never smoker Never smoker smoker Former smoker Current Unknown
Confidential: 1.00 (reference) to 1.69) 1.21 (0.87 (reference) 1.00 to 1.31) (0.85 1.05 to 1.33) (0.92 1.11 For to 2.59) (0.96 1.58 Review Only Supplementary Table 2. Adjusted HRs for the Association Between the Use of ACEI and the Risk of lung cancer of the lung Risk and theACEI of Use HRs Between for Association the 2.Table Adjusted Supplementary Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, Adjusted cohort sex,age,of for year
ARBs ACEIs cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia, (including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, Abbreviations: HR, hazard ratio; CI, confidence interval; ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, ofuse hypertension, treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry. before intheclasses year Page 35 of 42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 36 of 42 7 ‡ ‡ 1.12 (0.98 to 1.28) to 1.28) (0.98 1.12 (95% HR CI) Adjusted
Crude HR Crude 1.29 1.29 1.00 [Reference] [Reference] 1.00 [Reference] 1.00 * (95% CI) (95%
BMJ (lag: 2 years) 2 years) (lag: Incidence rate
https://mc.manuscriptcentral.com/bmj 2792 1,659,614 to 1.7) 1.7 (1.6 245 185,766 to 1.5) 1.3 (1.2 Confidential: For Review Only Events Person-years † Supplementary Table 3. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk ofCancerthe Lung Risk Useand theACEI of Between for Association HRs the Adjusted Table and Crude 3. Supplementary ARBs ACEIs Exposure Exposure Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, cohort sex,age,of for year Adjusted Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia,(including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, Abbreviations: HR, hazard ratio; CI, confidence interval; ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡
and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, ofuse hypertension, treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry before intheclasses year 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 8 ‡ ‡ Adjusted HR (95% (95% HRCI) Adjusted 1.17 (1.01 to 1.35) to1.35) (1.01 1.17
Crude HRCrude * BMJ (lag: 3 years) 3 years) (lag: (95% CI) (95% Incidence rate Incidence rate 1.7 (1.7 to 1.8) 1.7 (1.7 1.34 1.3 (1.1 to 1.5) 1.3 (1.1 1.00 [Reference] [Reference] 1.00 https://mc.manuscriptcentral.com/bmj
2396 1,375,795 Confidential:210 160,174 For Review Only Events Person-years † Supplementary Table 4. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk ofCancer the Lung Risk Useand theACEI of Between for Association HRs the Adjusted Table and Crude 4. Supplementary Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, cohort sex,age,of for year Adjusted ARBs ARBs ACEIs Exposure Exposure Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia, (including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, Abbreviations: HR, hazard ratio; CI, confidence interval; ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡
and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, ofuse hypertension, treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry. before intheclasses year Page 37 of 42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 38 of 42 9 ‡ ‡ 1.14 (1.00 to 1.29) to1.29) 1.14 (1.00 (95% HRCI) Adjusted
Crude HR Crude BMJ * (95% CI) (95% Incidence rate Incidence rate Cancer (Multiple Imputation) Imputation) Cancer(Multiple
https://mc.manuscriptcentral.com/bmj 3186 3186 1,977,139 (1.6 to 1.7) 1.6 1.32 266 266 213,557 (1.1 to 1.4) 1.2 [Reference] 1.00 1.00 [Reference] Events Person-years
Confidential:† For Review Only Supplementary Table 5. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk Lung Risk ofthe Useand ACEI of the Between for Association HRs the Adjusted and Crude 5.Table Supplementary Exposure Exposure ARBs ARBs ACEIs Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, cohort sex,age,of for year Adjusted Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia,(including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, Abbreviations: HR, hazard ratio; CI, confidence interval; ACEI, Angiotensin converting enzyme inhibitors; ARB, Angiotensin receptor blockers blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡
and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, ofuse hypertension, treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry. before intheclasses year 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 10
‡ Adjusted HR (95% (95% HRCI) Adjusted 1.20 (1.06 to 1.36) to1.36) (1.06 1.20
Crude HR Crude
entry.
BMJ * (Disease Risk Score) Score) (Disease Risk (95% CI) (95% Incidence rate Incidence https://mc.manuscriptcentral.com/bmj
3186 3186 1,977,139 to1.7) (1.6 1.6 1.32 266 266 213,557 to1.4) (1.1 1.2 [Reference] 1.00 [Reference] 1.00 Confidential: For Review Only Events Person-years † Supplementary Table 6. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk ofCancer the Lung Risk Useand theACEI of Between for Association HRs the Adjusted Table and Crude 6. Supplementary ARBs ACEIs Exposure Exposure Stratified on disease risk score decile. Variables included in the disease risk score included; age, sex, year of cohort entry, body mass index,body entry, year of cohort sex, age, score included; risk inthe disease included Variables scoredecile. risk on disease Stratified Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per
smoking, alcohol related disorders (including for example alcoholism, alcoholic cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), hepaticflexure), andalcoholic hepatitis cirrhosis the liver, of alcoholic for alcoholism, example (including disorders alcohol related smoking, treateddisease), duration of pulmonary chronicofobstructive pneumoniaandhistory (including entry diseasesto cohort of lung history prior cohort before inthe classesyear drug number uniqueof andthe total hypertension blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡ Page 39 of 42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 40 of 42 ‡ ‡ 11 Adjusted HR (95% (95% CI) HR Adjusted 1.26 (1.03 to 1.55) to1.55) (1.03 1.26
Crude HR Crude
* BMJ (95% CI) (95% Incidence rate Incidence 0.1 (0.1 to 0.1) to0.1) (0.1 0.1 1.44 0.1 (0.09 to 0.1) to0.1) (0.09 0.1 [Reference] 1.00 [Reference] 1.00 (Marginal Structural Model) Model) Structural (Marginal https://mc.manuscriptcentral.com/bmj
3186 3186 24,284,790 Confidential: 266 2,619,749 For Review Only Events Person-months † Supplementary Table 7. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk ofCancer the Lung Risk Useand theACEI of Between for Association HRs the Adjusted Table and Crude 7. Supplementary Adjusted for age, sex, year of cohort entry and the total number of unique drug classes in the year before cohort entry. Time updated variables Timeupdated entry. variables cohort classes year before in the drug of unique andtotaltheentry number cohort sex,age,of for year Adjusted Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Months. Person Months. 1000 Per ARBs ACEIs Exposure hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, and history of chronic and history tuberculosis, pneumonia, (including cohort entry diseasesto lung of prior history flexure), and hepatic hepatitis statins. of anduse hypertension treated durationof disease), obstructive pulmonary blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡ included body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic cirrhosis of the liver, alcoholic liver, of the cirrhosis alcoholic alcoholism,example for (including alcohol related disorders smoking, mass index, body included 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 12 ‡ 1.06 (1.00 to 1.13) 1.13) to 1.06 (1.00 1.14 (1.04 to 1.24) to1.24) 1.14 (1.04 1.03 (0.96 to 1.10) 1.10) to 1.03 (0.96 1.23 (1.05 to 1.45) to1.45) 1.23 (1.05 (95% CI) (95% Adjusted HR Adjusted
1.22 1.22 1.06 1.06 1.38 1.38 Crude HR Crude 1.00 [Reference] [Reference] 1.00 1.00 [Reference]
*
(95% CI) (95%
Incidence rate Incidence
BMJ
https://mc.manuscriptcentral.com/bmj 1511 1511 1,074,713 to1.5) (1.3 1.4 3186 3186 1,977,139 to1.7) (1.6 1.6 1.12 Events Events Person-years 905 905 457,309 to2.1) (1.9 2.0 2084 2084 197 1,440,232 to1.5) (1.4 1.4 79,598 to2.8) (2.1 2.5 (Using Thiazide Diuretic Use as an Alternative Comparator Group) Comparator Alternative Thiazide Diuretic as Use an (Using
Confidential: For Review Only
† Supplementary Table 8. Crude and Adjusted HRs for the Association Between the Use of ACEI and the Risk ofCancer the Lung Risk Useand theACEI of Between for Association HRs the Adjusted Table and Crude 8. Supplementary Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, cohort sex,age,of for year Adjusted Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per ≤ 5 5 ≤ 5.1 10 10 > cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia,(including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡ and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, use hypertension, of treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry. before intheclasses year Cumulative Duration of ACEI use (years) (years) ofuse ACEI Duration Cumulative Thiazide diuretics Thiazide diuretics ACEIs Exposure Exposure Page 41 of 42 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 42 of 13
‡ 0.98 (0.84 to 1.15) to 1.15) (0.84 0.98 0.87 (0.69 to 1.09) to 1.09) (0.69 0.87 0.85 (0.51 to 1.41) 1.41) to (0.51 0.85 (95% CI) (95% Adjusted HR HR Adjusted 0.94 (0.82 to 1.07) to 1.07) (0.82 0.94
Crude HR Crude
* BMJ (95% CI) (95% Incidence rate
Lung Cancer Lung
https://mc.manuscriptcentral.com/bmj 1511 1,074,713 to 1.5) 1.4 (1.3 [Reference] 1.00 [Reference] 1.00 173 144,546 to 1.4) 1.2 (1.0 0.88 78 59,047 to 1.6) 1.3 (1.0 0.79 15 9964 to 2.5) 1.5 (0.8 0.80 Events Person-years 266 213,557 to 1.4) 1.2 (1.1 0.85
Confidential: For Review Only
† Adjusted for age, sex, year of cohort entry, body mass index, smoking, alcohol related disorders (including for example alcoholism, alcoholic alcoholic alcoholism, example for (including disorders alcohol related smoking, mass index, body entry, cohort sex,age,of for year Adjusted Use of other antihypertensive drugs is considered in the model, but not presented in the table. thetable. in presented is in thenot but model, considered drugs antihypertensive ofUse other Per 1000 Person Years. Person Years. 1000 Per cirrhosis of the liver, alcoholic hepatitis and hepatic flexure), history of lung diseases prior to cohort entry (including pneumonia, tuberculosis, pneumonia, (including tuberculosis, entry cohort diseasesto prior lung of history flexure), and hepatic alcoholichepatitis of cirrhosis the liver, blockers receptor ARB, inhibitors; Angiotensin enzyme converting Angiotensin ACEI, interval; confidence ratio;CI, hazard HR, Abbreviations: * † ‡ and history of chronic obstructive pulmonary disease), duration of treated hypertension, use of statins, and the total number of unique drug drug uniqueof andnumber thetotal statins, ofuse hypertension, treatedof disease), duration pulmonary ofobstructive chronic andhistory cohort entry before intheclasses year > 10 10 > ≤ 5 5 ≤ 5.1 10 5.1 10 Supplementary Table 9. Crude and Adjusted HRs for the Association Between the Use of ARBs Compared to Thiazide Diuretics and the Risk of Risk the Diuretics and Thiazide to ofUseCompared the ARBs Between HRs for Adjusted Association the and 9.Table Crude Supplementary Exposure Cumulative duration ofuse (years) ARB duration Cumulative Thiazide diuretics Thiazidediuretics ARBs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60