Table 4.
Characteristics of included studies (Part I: study population and study type).
| Study | Study population |
Study type |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Study ID | Total number of study participants | Number of female study participants | Country of study population | Geographic location (specify as 'national' or list regions or sites) | Industrial sector (specify ISIC-4 code provided in worksheet “Industrial sector codes”) | Occupation (specify ISCO-08 code provided in worksheet “Occupation code”) | Age | Study design | Study period (month of first collection of any data and month of last collection of any data) | Follow-up period (period in months between exposure and outcome) |
| Chang et al. (2013) | 578 | 0 | Taiwan | Local | 30 | 7232 | 27.7 ± 5.3 years | Cohort study (prospective) | 1998–2008 | 9.8 ± 5.2 years |
| Davies (2002) | 27,499 (this is the analysis sample) | 0 | Canada | Local | 16 | 8172 | Mean 29,7 years, range 10,6–76,3 years | Cohort study (retrospective) | 1950–1995 | Mean 24,3 years, range 1–46 years |
| Eriksson et al. (2018b) | Baseline-5,753 | 0 | Sweden | Local | Unclear | Unclear | Baseline 55.3 (2.1) years, range 50–59 years | Cohort study (prospective) | 1974–1977; 2004 (last noise exposure data) | Unclear |
| Girard et al. (2015)* | 644 | 0 | Canada | Regional | 25 | Unclear | 55–64 years (cases mean = 60.0, controls mean = 58.8) | Case-control | 1983–2005/07 | Cases - mean: 31.9 years, controls: 29.8 years, all study subjects: 30.5 years |
| Gopinath et al. (2011) | Blue Mountains Eye Study (BMES-1) 1992–4 – 3,654 participants BEMS-2 (1997–1999) −3,509 participants; BEMS-3 (2002–2004)-1,952 participants | Unclear (BMES-1) 1992–4 – controls: 1,348 females, Exposed: 306 females BMES −2 –no data about females BMES-3 1,556 participants-917 females |
Australia | National | No data | No data | 67.9 ± 9.4 years (unexposed group) and 67.1 ± 8.9 (exposed) | Cohort study | Baseline: 1997–1999 -incidence study, 2002–2004 - mortality study, cut off point for CHD and stroke death -end of December 2007 | Prevalence data was obtained from BMES-2 (baseline), while, incidence analyses used data obtained from both BMES-2 and 5-year follow-up examination (BMES-3). |
| Huo Yung Kai et al. (2018)* | 1,156 | About 547 | France | National | No data | No data | 32 years, 42 years, 52–62 years |
Cohort study (prospective) | 2001–2006 | 5 years |
| Ising et al. (1997) | 2,543 | 0 | Germany | Local | No data | No data | 31–70 years | Case-control study | Unclear | No data |
| Kersten and Backe (2015) | 4,113 | 1,059 | Germany | Regional | Unclear | 1111, 1112, 1321, 1324, 1330, 1420, 2149, 2263, 2269, 2351, 2359, 2433, 4221, 4323, 6113, 7223, 7233, 7322, 7549, 8111, 8121, 9214, 9216 | 20–69 years | Case-control study |
No data | N/A |
| McNamee et al. (2006)* | 1,220 | 0 | United Kingdom | Regional | 35 | 7131 |
≤75 years | Case-control | 1965–1998 | ≤1 month to ≥40 years |
| Pettersson et al. (2020) | 166,088 (analysis sample) | 0 | Sweden | National | 41–43 | No data | 15–67 years | Cohort study | 1971–1993 | 17–40 years (ended in 2010) |
| Song (2013)* | 221 cases and 1,105 controls | 106 cases and 530 controls | Canada | National | 01 | 1221 | < 30 to > 55 years | Case-control | 31.12.2001–31.12.2009 | 180 |
|
Stokholm et al. (2013a) |
145,190 | 36,788 | Denmark | Regional | 1–4; 7–9 | 8160, 7322, 8112, 8121,8122, 8211, 7231, 8172, 1323, 4419, 7549,8219, 1120 | <25 years 25–34 years 35–44 years 45–54 years 55–64 years ≥65 years |
Cohort study (prospective) | 2001–2007 | 7 years |
| Stokholm et al. (2013b) | 164,247 | Unclear | Denmark | Regional | 1–4; 7–9 | 8160, 7322, 8112, 8121,8122, 8211, 7231, 8172, 1323, 4419, 7549,8219, 1120 | Unclear | Cohort study (prospective) | 2001–2007 | 7 years |
| Suadicani et al. (2012) | 5,249 (in 1970) 3,387 (in 1985–1986) |
0 | Denmark | National | 32, 43 | 3115, 8211, 8121, 8311, 2163, 5112 | 62.7 (5.2) years | A follow-up study to a cross-sectional survey | 1970–1986 | 16 years |
| Tessier-Sherman et al. (2017) | 2,052 | 0 | USA | Unclear | Section B Mining and quarrying: 24 Manufacture of basic metals, 32 Other manufacturing | 8121 | Mean 35.8, SD 8.5 | Cohort study (retrospective) | After 1 January 1996 to 31 December 2012 | 72 months; follow-up time - mean 6.5 years |
| Tong et al. (2017)* | 935 | 0 | China | Regional | Iron and steel enterprise (cold rolling and gas factory) | Unclear | ≤ 55 years, Essential Hypertension Group − 38.44 ± 8.51 years; Control Group − 38.11 ± 8.04 years | Case-control | February 2014 to July 2014 | No follow-up |
| Virkkunen et al. (2005) | 6,005 | 0 | Finland | National | Iron and metal work, machine work in plants, woodworking, and chemical process work | Unclear | 40–56 years at entry | Cohort study (prospective) | 1982–1999 | 15.9 years |
| Study |
Exposure assessment |
Comparator |
||||||
|---|---|---|---|---|---|---|---|---|
| Study ID | Exposure definition (i.e. how was the exposure defined?) | Unit for which exposure was assessed | Mode of exposure data collection | Exposure assessment methods | Levels/intensity of exposure (specify unit) | Number of study participants in exposed group | Number of study participants in unexposed group | Definition of comparator (define comparator group, including specific level of exposure) |
| Chang et al. (2013) | 8-hour time-weighted average equivalent sound level with and without adjustment for usage of HPDs (in dBA) | Individual | Technical device | Measurements and questionnaire on HPDs use | < 80 dBA; 80-85 dBA; ≥85 dBA (used in our analyses) |
205 (< 80 dBA) 221 (80 to <85 dBA) 152 (≥85 dBA) |
205 | < 80 dBA (low exposure group) |
| Davies (2002) | Duration of exposure to noise levels exceeding a specific threshold in Leq dBA (used in meta-analysis); and Cumulative exposure in dBA-year | Individual | Historical exposure levels were estimated by a determinants of exposure regression model, developed using 1,900 personal dosimetry measurements | A combination of measurements, interviews, hygienists' assessment and modelling | For duration of exposure: <3 years (reference), 3–10 years, 10–20 years, 20–30 years, > 30 years for the thresholds > 85 dBA; (≥85 dBA for > 3 years was used for our analyses) |
N/A | N/A | Exposure to <85 dBA for < 3 years |
| Eriksson et al. (2018b) | To assess occupational noise exposure, a previously developed job-exposure matrix was applied | Group level: 129 unique job families | Job-exposure matrix | Measurements reports | <75dBA; 75–85 dBA; >85 dBA (used in our analyses) |
2,823 | 2,930 | Exposure to noise: medium < 75 dBA |
| Girard et al. (2015)* | Exposure to ≥ 90 dBA | Individual | Technical device | Measurements | Exposure to ≥ 90 dBA for < 27, 27–36.4, ≥ 36.5 years; Noise levels ≥90 dBA/8h, cases − 46%, control − 50.9% (used in our analyses) | 320 | 324 | < 90 dBA/8h |
| Gopinath et al. (2011) | Questionnaires on workplace noise exposure history | Individual | Questionnaire | Self-reported | Self-reported exposed status; duration of exposure: 0 years, <1–5 years, >5 years; severity of exposure: none, tolerable, unable to hear speech (used in our analyses) | 2,796 | 1,859 | Answer “No” to the question: “Have you ever worked in the noisy industry or noisy farm environment?” |
| Huo Yung Kai et al. (2018)* | The questions used were similar to those used in the 5th European survey on working conditions in the ESTEV study and in the previous VISAT articles |
Individual | Data from French prospective VISAT study | Self-reported | Exposed at baseline or in the preceding five years to (cannot hear a person who is 2–3 m away even if talking loudly) |
483 | 673 | Answer “No” to a question on occupational exposure to “loud” noise |
| Ising et al. (1997) | Subjective noise categories: 1+2 Refrigerator and typewriter 3. electric lawn-mower 4. electric drill 5. pneumatic drill |
Work noise level measured as 1-min mean level in relation to the subjective work noise category | Subjective evaluation of noise loudness based on questionnaire | Self-reported and objective measurement in the sample of 80 men using Norsonic Type 110 | Subjective noise categories Lower categories (1+2), higher categories (3+4+5) - these noise categories correspond to the median (25 percentiles) of LAeq,T>70 dBA. |
395 | 2,148 | Low-noise-exposed workers (noise categories: refrigerator+typewriter) |
| Kersten and Backe (2015) | Occupational noise (LEX,8h,subj) and (LEX,8h,obj) >55 dBA | Individual | Questionnaire, technical device, and experts judgements | Self-reported vocal effort and equipment catalogue specifications | 46–61 dBA, 62–84 dBA, 85–94 dBA, 95–124 dBA |
1,880 | 2,233 | 42–61 dBA |
| McNamee et al. (2006)* | Mean daily noise expo-sure level with adjustment for usage of HPDs (LEP,d in dBA; number (N) of years with LEP,d> 85 dBA; noise emmission level NIL (NIL=LEP,d +10×log N) |
Individual | Experts judgements based on company work histories and noise survey records | Extrapolation | Unexposed, <85 dBA; >1 year exposed to >85 dBA |
Total −1402, cases − 717, control − 685 |
Total −800, cases − 384, control − 416 |
<85 dBA |
| Pettersson et al. (2020) | Noise exposure was defined on a job exposure matrix | Group level: a noise exposure category was assigned for each working group in the cohort | Job-exposure matrix | Survey of working conditions carried out by industrial hygienists |
≤ 85 dBA; >85 dBA (after re-calculation by authors) | 54,480 | 111,608 | ≤85 dBA |
| Song (2013)* | Cumulative noise exposure (dBA-years) | Individual level | Job-exposure matrix | Job-exposure matrix and record linkage | < 85; 85–95; > 95 dBA-years | Cases/controls: 69/347 (85–95 dB); 76/419 (> 95 dB) | Cases/controls: 76/339 (<85 dBA) | <85 dBA-years |
| Stokholm et al. (2013a) | Mean, full-shift noise exposure levels (LAeq values in dBA) + cumulative exposure |
Individual | Technical device | Measurements | < 70 dB; > 80 dBA for <3, 3–9, 10–19, and ≥20 years; > 80 dBA for <3, 3–9 (used in our analyses), 10–19, and ≥20 years |
87,959 men, 15,728 women | 20,443 men, 21,060 women |
< 70 dB |
| Stokholm et al. (2013b) | Mean, full-shift noise exposure levels (LAEq values in dBA | Individual | Technical device | Measurements and extrapolation | < 70 dB; > 80 dBA for <3, 3–9, 10–19, and ≥20 years; > 80 dBA for <3, 3–9 (used in our analyses), 10–19, and ≥20 years |
496,036 | 425,763 | < 70 dB |
| Suadicani et al. (2012) | Exposure to noise at a level where it is necessary to raise voice | Individual level | Questionnaire | Self-reported vocal effort | Exposure to “loud” noise for > 1 years | 2,998 workers | 1,890 workers, noise level 0 years |
0 years of exposure to “loud” noise |
| Tessier-Sherman et al. (2017) | Exposures ever equal or exceed an 8-h time-weighted average | Individual level | Technical device, personal dosimetry measurements | Dosimetry | <82 dBA (referent); 82–84 dBA; 85–87 dBA; >88 dBA (> 82 dBA combined for our analyses) |
1,102 | 950 | Occupational exposure to noise <82 dBA |
| Tong et al. (2017)* | 1) The 40-hour time-weighted average (TWA) sound level, in dBA, 2) A cumulative noise exposure (CNE), in dBA x year (dBA-year) | Individual level | Technical device | Cumulative noise exposure (CNE) was determined taking into account noise levels and the years of noise exposure; time-weighted average according the type of work, detention time, and work shift situation (used in our analyses) |
<85 dBA; ≥85 dBA time-weighted average (used for our analyses) | 461 | 474 | <85 dBA time-weighted average |
| Virkkunen et al. (2005) | Exposure to continuous noise (used in our analysis), exposure to impulse noise & continuous noise | Individual | Job-exposure matrix | Job-exposure matrix and record linkage | < 80 dBA; 80-85 dBA; >85 dBA dBA |
2,893 | 3,556 | < 80 dBA |
| Study |
Outcome assessment |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| Study ID | Definition of outcome | Which International Classification of Diseases (ICD) code was reported for the outcome (if any)? | Method of outcome assessment | Diagnostic assessment method | Specification of outcome | Number of cases with outcome of interest in exposed group | Number of non-cases (i.e. without outcome of interest) in exposed group | Number of cases with outcome of interest in unexposed group | Number of non-cases (i.e. without outcome of interest) in unexposed group |
| Chang et al. (2013) | Hypertension | None | Questionnaire, Blood pressure measurements | Self-reported diagnosed hypertension or SPB≥ 140 mmHg and/or DBP ≥90 mmHg | Incident hypertension | 141 | 437 | 44 | 161 |
| Davies (2002) | Hypertensive heart disease; ischaemic heart disease (IHD); acute myocardial infarction; stroke mortality | Hypertensive diseases (ICD9 401–405.9); IHD (ICD9 411–414.9, 429.2); acute myocardial infarction (ICD 410–410.9); stroke (cerebrovascular disease, ICD9 430–438.9) | Death certificate | Administrative record | Hypertensive heart disease; ischaemic heart disease; acute myocardial infarction; stroke mortality | In the groups > 3 years: hypertensive heart disease (n = 22), IHD (n = 693), acute MI (n = 757), stroke (n = 325) | Unclear | In the reference group < 3 years: hypertensive heart disease (n = 4), IHD (n = 123), acute MI (n = 153), stroke (n = 48) | Unclear |
| Eriksson et al. (2018b) | Coronary heart disease and stroke | ICD-8, ICD-9, ICD-10. CHD-410–414 (ICD-8, 9), and 120–125 (ICD-10); acute myocardial infarction 410 and 121; stroke 431–438 (ICD −8,9), 161–169 (ICD-10) | Hospital discharge national register | Hospital discharge national register | CHD, stroke, MI | CHD-medium noise 453, high noise 71; Stroke- medium noise 220, high noise 35 |
CHD - medium noise 2014; high noise −285 Stroke medium noise −2247, high noise −321, |
CHD − 480, stroke-262 | CHD 2450 Stroke 2668 |
| Girard et al. (2015)* | CVD mortality | ICD-9: 410, ICD-9: 411–414 + 429.2), CI M9 390–405; 415–459 (except 429.2) | Death certificate | Administrative record | Incident CVD mortality | 74 (exposed cases) | 0 (exposed cases) | 87 (unexposed cases) | 0 (unexposed cases) |
| Gopinath et al. (2011) | Angina, acute myocardial infarction, stroke | ICD − 9] codes 410.0, 411.0–8, 412, 414.0–9 and ICD- 10 (121.0–9, 122.0–9, 123.0–8, 124.0–9, 125.0–9, ICD −9: 430.0–438.9 and ICD-10160.0–169.9) | Medical history of participants, Australian National Death Index | Unclear | Prevalence/incidence of angina, acute myocardial infarction, stroke | Angina − 126 13.8%), AMI-98 (10.7%), stroke − 38 (4.1%), all CVD-171 (18.2%) | 675 | Angina − 168 (9.2%), AMI-115 (6.4%), stroke − 80 (4.4%), all CVD −218 (17.7%) | 1496 |
| Huo Yung Kai et al. (2018)* | Hypertension | None | BP was measured using an automatic standard sphygmomanometer (OMRON 705CP) |
SBP ≥ 140 mmHg and/or a DBP ≥ 90 mmHg and/or taking a antihypertensive medication |
Hypertension | 26 | 99 | 108 | 542 |
| Ising et al. (1997) | Myocardial infarction | ICD 410 | Hospital discharge record | Hospital discharge record | Myocardial infarction | 246 |
927 | 149 | 1221 |
| Kersten and Backe (2015) | Myocardial infarction | None | Computer assisted standardized interview | Physician diagnostic record | Myocardial infarction | 166 | 199 | 1493 | 1658 |
| McNamee et al. (2006)* | IHD mortality | ICD-9: 410–414 | Death certificate | Administrative record | Incident ICD mortality | 717 (exposed cases) |
685 (unexposed control) |
384 (unexposed cases) | 416 (unexposed control) |
| Pettersson et al. (2020) | Myocardial infarction and stroke | IHD: ICD-8410–412, ICD- 9410–412, and ICD-10I21-I25; Stroke: ICD-8430–438, ICD-9430–438, and ICD-10I60-I69 |
National Cause of Death Register | Administrative record | Myocardial infarction and stroke | Myocardial infarction: 1,943 Stroke: 534 |
Myocardial infarction: 52,537 Stroke: 53,946 |
Myocardial infarction: 4,164 Stroke: 1,116 |
Myocardial infarction: 107,444 Stroke: 110,492 |
| Song (2013)* | CVD | None | Questionnaire | Self-reported heart disease | Positive response | 64 (85–95 dB) /78 (> 95 dB) | 331 (85–95 dB) /419 (> 95 dB) | 76 | 339 |
| Stokholm et al. (2013a) | Hypertension | ICD-8 codes, ICD-10 codes, but exact codes uncelar | Data on redeemed anti-hypertensive prescription, hospital discharge | Administrative record | Incidence of hypertension /1000 person-year | Men 6,051 Women 1,603 |
Men 81,908 Women- 19,457 |
Men 1,536, Women 2,205 |
Men 18,907, Women 18,855 |
| Stokholm et al. (2013b) | Stroke | DI61, DI63 D164 |
Danish National Patient Register | Unclear | Incident stroke | 638 | Unclear | 343 | Unclear |
| Suadicani et al. (2012) | IHD mortality | IHD codes 410–412, ICD (1994) 120–125 | Danish National Civil Registry | Physician diagnoses in national registry | IHD mortality | 197 deaths due to IHD | 2,801 | 6.4% of 1890 subjects | 93.6% |
| Tessier-Sherman et al. (2017) | Hypertension | ICD9, 401–404 | Central data processing vendor for all employees | Administrative datasets | Hypertension | 244 | 1,808 | No data | No data |
| Tong et al. (2017)* | Hypertension | None | Physical examination | Physician diagnostic record | Hypertension | 182 | 279 | 130 | 344 |
| Virkkunen et al. (2005) | Coronary heart disease | CHD - codes 410–414 in the ninth revision of the ICD and I20-I25 in the tenth revision | CHD end points were obtained from official Finnish registers | Hospital discharge record | Coronary heart disease | 515 | 2378 | 509 | 3047 |
| Study |
Adjustments of effect estimates in model prioritized by reviewers |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Study ID | Adjusted for confounding by: age | Adjusted for confounding by: sex | Adjusted for confounding by: Socioeconomic status (please specify indicator, e.g. level of education) | Other potential confounders adjusted for (please specify) | Adjusted for mediation by: tobacco smoking | Adjusted for mediation by: Alcohol use | Adjusted for mediation by: obesity | Other potential mediators adjusted for | Interactions adjusted for | Adjustment for clustering (if any) |
| Chang et al. (2013) | Yes | N/A (males only) | Educational level | Body mass index, employment duration, cigarette use, alcohol intake, exercise | Yes | Yes | Yes | No | No | No |
| Davies (2002) | Yes | N/A (males only) | No | Calendar year and race | No | No | No | No | No | No |
| Eriksson et al. (2018b) | Yes | N/A (males only) | No | No | No | No | No | No | Interaction between occupational noise and high strain | No |
| Girard et al. (2015)* | Yes | N/A (males only) | No | No | No | No | No | No | No | No |
| Gopinath et al. (2011) | Yes | Yes | Occupational prestige | Body mass index, smoking, walking difficulties and self-reported poor health | Yes | No | Yes (stroke incidence model) | Yes | No | No |
| Huo Yung Kai et al. (2018)* | Yes | Yes | Educational attainment | Body mass index, smoking habits, daily alcohol intake, leisure time physical activity, history of diabetes, history of hypercholesterolemia, treatment for hypertension, working status and initial blood pressure | No | No | No | Yes | No | No |
| Ising et al. (1997) | Yes | N/A (males only) | Social class, education, marital status, housing area | Body mass index, Social class, Education, Marital status, residential area, shift work, Current smoking | Yes | No | Yes | No | No | No |
| Kersten and Backe (2015) | Yes (matching variable) | Yes (matching variable) | Current employment status, <12 years at school | Shift work, work >40h per week | No | No | No | No | No | No |
| McNamee et al. (2006)* | Yes | N/A (males only) | No | Pre-employment measures and duration of employment | No | No | No | No | No | No |
| Pettersson et al. (2020) | Yes | N/A (males only) | No | Region | No | No | No | No | No | Yes |
| Song (2013)* | Yes (matching variable) | Yes (matching variable) | Education, family income | Smoking, body mass index, drinking, smoking, physical activity, hypertension | Yes | Yes | Yes | Yes | No | No |
| Stokholm et al. (2013a) | Yes | Yes | Five categories, blue/white collar | Calendar year, employment status | No | No | No | No | Interaction between sex and occupation | Yes |
| Stokholm et al. (2013b) | Yes | Yes | Socioeconomic status | Calendar year, employment status | No | No | No | No | No | No |
| Suadicani et al. (2012) | Yes | N/A (males only) | Low social class | Physical activity, cumulative tobacco consumption, alcohol intake | Yes | Yes | Yes | No | Age + lifestyle and social class, age + clinical factors, age + all potential confounders | No |
| Tessier-Sherman et al. (2017) | Yes | N/A (males only) | Economic status, job category, annual wages | Body mass index, smoking, hearing acuity | Yes | No | Yes | Yes | No | Yes |
| Tong et al. (2017)* | No | N/A (males only) | No | Body mass index, low density lipoprotein cholesterol, hypertension family history, A1166C gene | Yes | No | No | Yes | No | No |
| Virkkunen et al. (2005) | Yes | N/A (males only) | No | Systolic blood pressure | No | No | No | Yes | No | No |
| Study |
Prioritized model |
Estimate of effect of exposure on outcome |
|||
|---|---|---|---|---|---|
| Study ID | Are two or more alternative models reported? | Which of the alternative models was prioritized/selected for use in the review and/or meta-analysis? | Reason for prioritization/selection | Treatment effect measure type | Was an exposure–response (or dose–response) analysis conducted? |
| Chang et al. (2013) | Yes | Relationships between noise exposure and hypertension in total | N/A | Hazard ratio | No |
| Davies (2002) | Yes | The model yielding RR of different cardiovascular outcomes in those exposed to > 85 dBA for >3 years vs. exposed to >85 dBA for <3 years | This duration of exposure was most biologically plausible, as exposed for <3 years would be unlikely to cause cardiovascular disease | Relative risk | Trend per increasing duration of exposure (not of interest for pooling) |
| Eriksson et al. (2018b) | Yes – age-adjusted and fully-adjusted model (body mass index, diabetes, hypertension, smoking, cholesterol) | Hazard ratio adjusted for age only | Overadjustment for potential mediators in the fully-adjusted model | Hazard ratio | No |
| Girard et al. (2015)* | Yes – models for duration of noise exposure and crude 2x2 table | Raw data in descriptive | The duration of exposure categories are not comparable to the exposure categories in other studies | Calculated relative risk | No |
| Gopinath et al. (2011) | Yes – incidence and mortality | Only the mortality model, because for the incidence model, the only significant effect was selectively reported, and it was based on only 4 cases with stroke | N/A | Hazard ratio and Odds ratio | Trend per increasing duration of exposure (not of interest for pooling) |
| Huo Yung Kai et al. (2018)* | Yes, crude and adjusted models (age, gender, body mass index, smoking, alcohol, physical activity, diabetes, hypercholesterolemia, employment, educational attainment) | Unadjusted model (calculated from raw data) due to gross adjustment for mediators in the adjusted model; Moreover, adjusted estimate was reported only in a Fig. with poor resolution | The adjusted models revealed that most of these associations were explained by individual cardiovascular factors, except for the negative effect of high job strain and positive effect of job recognition which had an independent role | Odds ratio reported, but we calculated relative risk from raw data | No |
| Ising et al. (1997) | Yes – crude and adjusted models | Model adjusted for smoking, body mass index, age, social class, education, marital status, shift work, housing area | Control for confounding factors with acceptable adjustment for potential mediators | Odds ratio | No |
| Kersten and Backe (2015) | Yes – models using all occupational groups and stratified by occupational group | The model using all occupational groups combined (for men and women) | Insufficient number of cases in the stratified models | Odds ratio | No |
| McNamee et al. (2006)* | Yes – crude and two adjusted; data from both sampling sites vs. data from one site | Adjusted model taking into account both sites | Control for confounding factors and acceptable adjustments for potential mediators; Moreover, the estimates do not differ between crude and adjusted models | Odds ratio | Yes |
| Pettersson et al. (2020) | Yes – re-calculated upon request | Relative risk adjusted for age and region | Parsimony | Relative risk | No |
| Song (2013)* | Yes – crude and adjusted models | Multivariate logistic regression | Control for confounding factors and acceptable adjustments for potential mediators | Odds ratio | No |
| Stokholm et al. (2013a) | Yes – crude and adjusted models | Adjusted model | Control for confounding factors | Relative risk | No |
| Stokholm et al. (2013b) | Yes – crude and adjusted models | Adjusted model | Control for confounding factors | Hazard ratio | Trend RR by 1-unit dBA-year increase (not of interest for pooling) |
| Suadicani et al. (2012) | Yes – four models | Age + lifestyle and social class-adjusted model | This model seems the best compromise between parsimony and controlling for confounders | Hazard ratio | No |
| Tessier-Sherman et al. (2017) | Yes – crude and adjusted models | Model adjusted for age, body mass index, smoking | Control for confounding factors and acceptable adjustments for potential mediators | Relative risk | Yes |
| Tong et al. (2017)* | Yes – models for time-weighted average and cumulative noise exposure | Time-weighted average model | Allows comparison with the other studies that used this noise metric | Odds ratio | No |
| Virkkunen et al. (2005) | Yes – different follow-up models and estimates for continuous and impulse noise | The longest follow-up model and continuous noise | The other follow-ups yield similar effect estimates; relatively few workers are exposed to impulse noise | Relative risk | Trend per increasing level of exposure (could not be pooled) |
*The study provided deprioritized evidence and was not included in the main meta-analysis due to it being a single case-control study in the respective model (Girard et al., 2015, McNamee et al., 2006, Tong et al., 2017), unadjusted estimate extracted (Huo Yung Kai et al., 2018) or incomparable noise metric (Song, 2013).