Illustrations v.5.9 - Additional restrictions

Author

Andrei Pyko

Published

September 12, 2024

Task:

Exposure-response curves, inflections point (input/suggestions from Timo/Pekka)

  1. Restrict curves to noise levels 75 dB or below (only few above this exposure – Mette checked in Dk population). Not performed yet but Andrei expects that it will not affect results.

  2. Exclude studies with very high “lower level” – at least in a sensitivity analysis (may have large effect on the shape of the exposure-response curve). Which lower level do we choose? Decide on 60 dB.

  3. CIs: very wide CI in the figures but narrower in the actual estimates. Why are the CIs wider in the curves compared to the estimates. Andrei – send IHD estimates to Pekka so he can check – mainly for us to understand why there is this difference. Andrei and Pekka will email regarding this. - on the way

  4. Now 40 dB in curves as reference point (CI on 1) – should we set this to median noise instead? We keep reference at 40 dB.

Material and Methods

This analysis was conducted using the most recent updates from June 2024. For all tests, we employed spline models to evaluate the exposure-response relationships, utilizing the dosresmeta package for meta-analytic dose-response models. The spline model used restricted cubic splines to flexibly model non-linear associations between noise exposure and health outcomes. Spline function with 4 degrees for freedom were used. The model was built with the common meta-analytic specifications such as study weights, variances, and covariates, which were extracted from the dataset using a standardized process.

The following restrictions and tests were performed in this analysis:

  1. Restriction to noise levels ≤ 75 dB: For each outcome, data above 75 dB were excluded, considering the potential outlier effect of higher exposure levels. Exclusion of studies with very high “lower levels”: Studies with baseline exposure levels greater than or equal to 60 dB were excluded from sensitivity analyses to assess their impact on the exposure-response curves.

  2. Comparison of confidence intervals (CIs): We examined why the confidence intervals appeared wider in the exposure-response curves compared to actual risk estimates.

  3. Results from the restricted and unrestricted models were visually compared through the generation of exposure-response curves, allowing for a direct assessment of the impact of these restrictions.

1. 75 db Restriction

For each outcome separately additional restriction was applied excluding observations above 75 dB. We take two linear models (unrestricted and restricted), along with a title and reference point as inputs. Then we predicts the relative risk (RR) values for a sequence of exposure levels, normalizes the RR values to the reference point, and plots the predictions with confidence intervals for both models.

IHD

Among IHD studies only Pyko et al., 2023 was affected by this restriction with original data, the rest of studies are affected in the highest extrapolated values (after “linearization”).

Studies with risk estimates at levels >= 75 dB
Study subtype Tranformed exposure, Lden Categorical RR (95% CI) adjusted for SES, lifestyle
Cai et al., 2018 // EPIC-OXFORD IHD 76.5 NA
Cai et al., 2018 // EPIC-OXFORD IHD 81.5 NA
Hao et al., 2022 IHD 75.3 NA
Hoffmann et al. 2015 IHD 79.1 NA
Pyko et al., 2023 // Recalculated IHD 75.0 Model 2: 40 dB: 1.00 (0.95-1.05), 45 dB: 1.00 (0.98-1.03); 50 dB: 0.99 (0.97-1.01); 55 dB: 1.00 (0.98-1.02); 60 dB: 1.02 (1.00-1.04); 65 dB: 1.04 (1.01-1.06); 70 dB: 1.08 (1.04-1.13); 75 dB: 1.16 (1.06-1.27) Size of studybase and number of cases in analysis provided in column AH.
Figure 1: Comparison of linear models for IHD studies with and without restriction to exposure levels <= 75 dB.

MI

In MI studies Babisch et al. 1994 was affected in two categories, Magnoni et al. 2021 and Pyko et al. 2023 with one category each.Hao et al. 2022 is effected in the highers extrapolated value only.

MI Studies with risk estimates at levels >= 75 dB
Study subtype Tranformed exposure, Lden Categorical RR (95% CI) adjusted for SES, lifestyle
Babisch et al., 1994 MI 75.0 <=60 dB: 1.0 (reference), 61-65 dB: 1.2 (0.8-1.7); 66-70 dB: 0.9 (0.6-1.4); 71-75dB: 1.1 (0.7-1.7); 76-80 dB: 1.5 (0.8-2.8) Number of cases and controls in analysis: 645 and 3390, respectively.
Babisch et al., 1994 MI 80.0 <=60 dB: 1.0 (reference), 61-65 dB: 1.2 (0.8-1.7); 66-70 dB: 0.9 (0.6-1.4); 71-75dB: 1.1 (0.7-1.7); 76-80 dB: 1.5 (0.8-2.8) Number of cases and controls in analysis: 645 and 3390, respectively.
Hao et al., 2022 MI 75.3 NA
Magnoni et al., 2021 MI 75.0 From Table 2. Lden < 65 dB: HR = 1 (ref); 65-69 dB: HR = 0.994 (0.951-1.040); 70-74 dB: HR = 1.005 (0.958-1.053); = 75 dB: HR = 0.999 (0.951-1.050). (N= 1,087,110)
Pyko et al., 2023 // Recalculated MI 75.0 Model 2: 40 db: 1.00 (0.92-1.08), 45 dB: 1.00 (0.95-1.04); 50 dB: 0.97 (0.94-1.01); 55 dB: 0.98 (0.95-1.01); 60 dB: 1.00 (0.97-1.03); 65 dB: 1.02 (0.97-1.06); 70 dB: 1.06 (0.99-1.14); 75 dB: 1.14 (0.99-1.31) Size of studybase and number of cases in analysis provided in column AH.
Figure 2: Comparison of spline models for MI studies with and without restriction to exposure levels <= 75 dB.

Stroke

In stroke non of the studies is affected by this restriction with orginal data but extrapolated linear associations.

Stroke Studies with risk estimates at levels >= 75 dB
Storke Studies with risk estimates at levels >= 75 dB
Study subtype Tranformed exposure, Lden Categorical RR (95% CI) adjusted for SES, lifestyle
Hao et al., 2022 Stroke 75.3 NA
Magnoni et al., 2021 Ischemic stroke 75.0 NA
Cai et al., 2018 // EPIC-OXFORD Cerebrovascular disease 76.5 NA
Cai et al., 2018 // EPIC-OXFORD Cerebrovascular disease 81.5 NA
Figure 3: Comparison of linear models for Stroke studies with and without restriction to exposure levels <= 75 dB.

2. Exclude studies with very high “lower level”

… at least in a sensitivity analysis (may have large effect on the shape of the exposure-response curve). Which lower level do we choose? Decide on 60 dB. Lden

Here is the table with all excluded studies for all outcomes because of high “lower level” (>= 60 dB Lden). No IHD studies are affected by this restriction.

Studies with high “lower level” (>= 60 dB Lden)
Study subtype Lden Original levels
Babisch et al., 1994 MI 62.0 <=60 db LAeq,6-22h: 1 (ref)
Babish et al., 2005 // Men MI 62.0 <=60 db LAeq,6-22h: 1.00 //
Babish et al., 2005 // Women MI 62.0 <=60 db LAeq,6-22h: 1.00 //
Carey et al., 2016 Stroke 64.1 <55 db Lnight: 1.00 (ref) /
Lekaviciute 2007 MI 63.3 <60 db LAeq,24h: 1.00 (ref)
Magnoni et al., 2021 Ischemic stroke 65.0 <65 db Lden: 1.00 (ref) //

Here are how that restriction affects exposure response curves for each of the outcomes one by one (“All studies” vs “Restricted sample” in figures legend). Beside to ERFs with and without the restrictions, original estimates of excluded studies are also presented (some of the CIs are ommited if they are outside of Y axis range, could be checked in main table 3 of the manuscript)

Stroke

MI