Background Philip Morris International, one of the largest transnational cigarette manufacturers, has heavily invested in its new heated tobacco product, IQOS, marketing it aggressively as a less harmful alternative to cigarette smoking. The company’s assertions that the product replaces cigarettes in a market have never been independently tested. The objective of this study is to determine whether introduction of IQOS affected cigarette sales in a large economy.

Data and Methods Using 2014 to 2018 monthly retailer panel data from Japan, we analyse whether different dates of IQOS introduction across Japan’s regions are reflected in the patterns of cigarette sales in those regions. A series of placebo models are estimated to test if events other than IQOS introduction could have better explained the observed trends in cigarette sales.

Results Cigarette sales begin to substantially decline at the time of the introduction of IQOS in each of 11 Japanese regions (Chow tests p<0.001). IQOS introduction, which varied across regions, better predicted the timing of cigarette sales decline than any one time applied to all regions simultaneously (a national-level exogenous shock) and than nearly all possible rearrangements of the true IQOS introduction months among the regions (exact permutation test’s p value from 0.02 to 0.13, depending on the study approach).

Conclusions The introduction of IQOS likely reduced cigarette sales in Japan. The net population health impact, however, cannot be assessed without resolving several key uncertainties related to the direct harms of IQOS and the precise patterns of both smoking and IQOS use.