Literature review | Long-Term Assessment | Sep 29, 2022

      Hospitalization rates for COPD and IHD started decreasing shortly after the launch of heated tobacco products in Japan

      In the absence of long-term data on smoking-related diseases, real-world data can be used to look for early signals of the population health impact of introduction of smoke-free products. In this real-world evidence study, we observed a statistically significant reduction in hospitalization rates for chronic obstructive pulmonary disease (COPD) and ischemic heart disease (IHD) in Japan after the introduction of heated tobacco products.  This article is an excerpt from Scientific Update Issue 16.

      Heated tobacco products (HTPs) are available in different markets. As observed in the Japanese market, the introduction of HTPs in stages from 2014 (the single city pilot) to 2016 (the national launch of PMI’s tobacco heating system, THS)* has accelerated the decline in cigarette sales without disrupting the continued decline of tobacco sales overall. Knowing this, it is important to take the next step and assess the impact of these products on the health of the individual and the population as a whole. So far, epidemiological study results are not available because the symptoms of smoking-related diseases can take decades to develop. In the absence of long-term epidemiological studies, real-world data can be used to look for early signals of the population health impact of introducing HTPs. 

      Read these independent studies that show that the decline in cigarette sales in Japan is likely caused by the introduction of HTPs:


      Declines in hospitalization for COPD exacerbation and IHD observed in early proof-of-concept real-world evidence study

      First, we conducted a proof-of-concept study to determine whether the introduction of HTPs in Japan could be correlated to hospitalization rates for COPD exacerbations or IHD. In fact, this early study did show a significant reduction in the rate of hospitalization due to COPD exacerbations after the introduction of HTPs in Japan. We also observed a small decline in the hospitalization rate for IHD, although not significant. 

      Image of the paper "Ischemic heart disease and chronic obstructive pulmonary disease hospitalizations in Japan before and after the introduction of a heated tobacco product

      Image of the first page of the article “Ischemic Heart Disease and Chronic Obstructive Pulmonary Disease Hospitalizations in Japan Before and After the Introduction of a Heated Tobacco Product.” This paper was published in Frontiers in Public Health in June 2022, by van der Plas et al.


      Real-world data source: The Japan Medical Data Center (JMDC) database

      A more comprehensive peer-reviewed study, published June 2022 in the journal Frontiers in Public Health, was conducted in accordance with the Guidelines for Good Epidemiological Practice (GEP). We replicated the earlier proof-of-concept study using a different real-world data source: the Japan Medical Data Center (JMDC) insurance claims database. The JMDC database contains accumulated receipts (inpatient, outpatient, and dispensing) and Diagnosis Procedure Combination (DPC) data of ~7.3 million patients (as of April 2020).

      Analyses were conducted using both all claims data (the broad definition), as well as only DPC data (the strict definition). DPC data are used in claims and in acute care hospitals in Japan. The DPC data are more detailed than the all claims data because it associates ICD codes directly to the reason for hospitalization. ICD means International Classification of Diseases.

      The study population consisted of adults (aged ≥20–74 years, employees of large corporations and dependents) who were hospitalized between January 2010 and December 2019. Analyses were done on the full-length data from 2010 through 2019 as well as a shorter period of 2013 through 2019. Because there was no significant difference between the two, the 2013-2019 data is shown.

      Among all hospitalizations, the selected endpoints for this study were:

      • monthly hospitalizations due to COPD (for all related International Classification of Diseases-10 codes),
      • COPD exacerbation,
      • COPD exacerbation plus lower respiratory tract infection (LRTI),
      • and IHD. 


      Interrupted time-series analysis reveals association between endpoints and intervention

      We conducted an interrupted time-series analysis to see whether there were changes in the endpoints listed above before versus after the intervention. For this study, the intervention is the introduction of HTPs to the Japanese market, and so the time of the intervention for modeling purposes was based on the level of market penetration of HTP sales. The market share of HTPs in Japan increased from 0.01% in January 2015 to about 7% in January 2017, and afterwards increased more rapidly to reach 25% by the end of 2019. For this reason, January 2017 was chosen for the starting point of the intervention.

      Besides the impact of time and the effects of HTP introduction, we also adjusted for the average age, sex distribution, seasonality, and annual rate of influenza vaccination.


      Statistically significant declines in hospitalization for COPD after introduction of HTPs in Japan

      We observed a statistically significant (P = 0.0001) reduction in the number of hospitalizations for COPD when using all available data. For COPD (all codes), hospitalization numbers dropped remarkably by 0.1–0.2% when comparing the pre- and post-HTP introduction time trends. 

      COPD (Broad) Hospitalization Rate Over Time JMDC Records

      COPD hospitalization rate over time JMDC hospital records

      This is figure 2 from van der Plas et al: Expected and observed trends in hospitalization numbers due to chronic obstructive pulmonary disease after introduction of heated tobacco products in Japan. 


      For COPD exacerbations there were few data points (only 179 in 2019), and the observed increase in hospitalizations was not statistically significant. For these reasons, results related to COPD exacerbations should be interpreted with caution.

      For COPD exacerbations plus LRTI, there was a decrease of 0.03-0.04% between the pre-and post-HTP introduction time trends, but otherwise an overall increasing trend. While such a decrease was observed both with and without adjustments for confounders, it was only statistically significant when confounders were not taken into account. Hospitalization numbers increased with average age and were also correlated with sex.


      Statistically significant declines in hospitalization for IHD after introduction of HTPs in Japan

      For hospitalizations due to IHD, the all claims data showed a lower proportion of hospitalizations during the post-HTP period compared to the pre-HTP years. Depending on which of the confounders were adjusted for, there was an upward spike in IHD hospitalizations at the time of HTP introduction, followed by an acceleration in the reduction of hospitalizations. However, the trend change was only statistically significant without adjusting for confounders when all data was considered. Again, hospitalizations increased with average age.

      When only DPC data for IHD was taken into account, the slight increase at the intervention point and also the decreasing trend after HTP introduction were statistically significant (P < 0.00001), with or without adjustments for confounders. 

      IHD Hospitalization Rate Over Time JMDC DPC Data

      .IHD hospitalization rate over time JMDC DPC data

      This is figure 3 from van der Plas et al. Expected and observed trends in hospitalization numbers due to ischemic heart disease after introduction of heated tobacco products in Japan, using only Diagnosis Procedure Combination data.



      The results of this time-trend analysis conducted using JMDC data demonstrate that there was a change in the trajectory of smoking-related disease hospitalization rates following the introduction of HTPs into the market. Specifically, we detected a significant reduction in the number of hospitalizations for COPD (all codes) when using all available data from the JMDC database, and for IHD when using the DPC claims data. We also found non-significant reductions in hospitalizations due to COPD exacerbation plus LRTIs and IHD after HTP introduction in Japan, when considering all data.

      This study does come with limitations. We considered age, sex, quarterly seasonality, and flu vaccination as potential confounders; we did not include additional factors influencing the hospitalization pattern, such as legislation and policy changes. Also, time-trend analyses like the one in this study do not assess causal relationships between exposure and outcome. They evaluate the potential impact of a population-based intervention on a population’s health.

      Analyses like these, however, are important in the context of epidemiological and public health research and can show fluctuations in the incidence of these and other noncommunicable diseases over many years.

      This study's findings provide insights into the potential impact of HTP commercialization on the hospitalizations associated with COPD and IHD. As a next step, these findings warrant further research to evaluate the impact of HTPs on the health of populations in other countries where these products have been introduced. Despite the study's limitations, findings from these studies provide important insights on the potential health impacts of HTPs before long-term epidemiological data becomes available.

      *THS is commercialized under the IQOS brand in various markets, including Japan.