Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. Correspondence Reduction in air in each of the Chinese cities, the bias 2 Petetin H, Bowdala D, Soret S, at al. Meteorology-normalised impact of COVID-19 could be even larger. pollution and Second, the baseline mortality lockdown upon NO₂ pollution in Spain. Atmos Chem Phys Discuss 2020; published attributable mortality used by the authors to quantify the online May 25. DOI:10.5194/acp-2020-446 (preprint). temporal reduction in number of due to COVID-19 deaths is from 2018. The choice of this 3 Vicedo-Cabrera AM, Sera F, Liu C, et al. Short term association between ozone and lockdown particular year is unjustified, and the mortality: global two stage time series study in 406 locations in 20 countries. BMJ 2020; use of data from one year can lead to published online Feb 10. DOI:10.1136/ Lockdown measures reducing the biases in the estimates, particularly for bmj.m108. 4 Liu C, Chen R, Sera F, et al. Ambient particulate transmission of COVID-19 have led to cause-specific mortality. In addition, air pollution and daily mortality in 652 cities. a temporal improvement in air quality Chen and colleagues1 used different N Engl J Med 2019; 381: 705–15. worldwide. An assessment by Kai Chen sources of data for the calibration of 5 Centre for Research on Energy and Clean Air. China’s air pollution overshoots pre-crisis and colleagues1 reported that a drop in historical exposure-response models levels for the first time. 2020. https:// air pollution in 367 Chinese cities was and the estimations of the reduction in energyandcleanair.org/wp/wp-content/ uploads/2020/05/China-air-pollution- associated with 8911 prevented deaths air pollution, which could lead to large rebound-final.pdf (accessed May 18, 2020). because of a reduction in nitrogen biases because of inconsistent choice dioxide (NO2; 95% CI 6950–10 866) of air pollution stations. and 3214 deaths from a reduction in Finally, the authors state that PM2·5 (95% CI 2340–4087). However, “the COVID-19 outbreak led to the method for quantifying reductions improvements in air quality that in air pollution and attributable brought health benefits in non- mortality is on the basis of several COVID-19 deaths”. We would like to simplifications and assumptions that raise a note of caution, given that are likely to bias the estimates. some of the prevented deaths might First, air pollution reductions correspond to the oldest and frailest estimated with the difference-in- individuals whose death was brought difference approach cannot be forward by just a few days or weeks, directly attributed to the lockdown or correspond to those individuals because this method does not cancel whose health is at risk given that air out the effect of weather on air pollution in China has already exceeded pollution. To show this point, we concentrations before the crisis.5 In our estimated the change in NO2 during opinion, improvements in air quality lockdown in 48 Spanish provinces, will result in a long-term benefit only if from March 15 to April 23, 2020, they are sustained in time. by estimating business-as-usual We declare no competing interests. concentrations with a meteorology Copyright © 2020 The Author(s). Published by normalisation technique based on Elsevier Ltd. This is an Open Access article under the machine learning,2 which shows the CC BY-NC-ND 4.0 license. expected concentrations that would Hicham Achebak, Hervé Petetin, have occurred without lockdown. Marcos Quijal-Zamorano, We then used exposure-response Dene Bowdalo, functions3,4 to transform the daily Carlos Pérez García-Pando, observed and business-as-usual Joan Ballester* NO 2 time series into daily NO 2- joan.ballester@isglobal.org attributable mortality time series, Centre for Demographic Studies, Autonomous and the difference was compared with University of Barcelona, Barcelona, Spain (HA); Climate and Health Program, Barcelona Institute for estimates obtained with the method Global Health, Barcelona 08003, Spain (HA, MQ-Z, in Chen and colleagues.1 Our unbiased JB); Earth Sciences Department, Barcelona method increased the reduction Supercomputing Center, Barcelona, Spain (HP, DB, CPG-P); and ICREA, Barcelona, Spain (CPG-P) in NO 2 by 11% on average and 1 Chen K, Wang M, Huang C, Kinney PL, increased the reduction in attributable Anastas PT. Air pollution reduction and deaths by 6·5%. Depending on how mortality benefit during the COVID-19 anomalous the weather was during outbreak in China. Lancet Planet Health 2020; 4: e210–12. lockdown and the reference periods www.thelancet.com/planetary-health Vol 4 July 2020 e268