Given the high number of analyses performed across subgroups and multiple ECG abnormalities studied, there is high risk of a statistically significant false positive finding. Minor ECG abnormalities are common in healthy black people and associated with physical activity; we might have not been able to fully adjust for that. Minor abnormalities have not been associated with future CVD in young adults. While these elements inherently limit our confidence in the measures of association, CARDIA is to our knowledge the only cohort with ECG data which assessed marijuana exposure over such a long time and in such a large cohort. While the study of daily marijuana users is of evident interest, few participants use marijuana daily, and we still need to know if marijuana use is associated with changes in ECG, even at lower exposures or from past exposure. The strength of the CARDIA dataset lies in the possibility to study the full spectrum of marijuana use intensity typical of the exposure in the general population. No other longitudinal study assessed the association between marijuana use and ECG. In this study, we did not find an association between cumulative or past use of marijuana and ECG abnormalities. Compared to the previous studies assessing the association between marijuana use and ECG, most of them published more than 4 decades ago and including only very few participants, we can now report results from a cohort more than 100 times larger than previous studies. We were able to adjust for a rich set of covariables repeatedly measured at 7 examinations and a dozen phone follow-ups. We cannot exclude the possibility of informative censoring, which is a source of bias,vertical farming equipment suppliers but addressed that issue by using IPAW, as stated in the methods section.The District of Columbia and 15 States – Alaska, Arizona, California, Colorado, Illinois, Maine, Massachusetts, Michigan, Montana, Nevada, New Jersey, Oregon, South Dakota, Vermont, and Washington have legalized recreational marijuana use.

As a result, involuntary exposure to secondhand marijuana smoke has become much more common in everyday settings across the country. Studies have shown that secondhand exposure close to tobacco smoking or vaping is substantially higher than farther away – this “proximity effect” will also be an issue near marijuana smoking or vaping. The initial research investigating the proximity effect and spatial variation of exposure near a source used a tracer gas to mimic the transport of emitted air pollutants. For example, McBride et al released carbon monoxide as a tracer in a residential living room while using 12 real time CO monitors to measure concentrations at different indoor positions. Acevedo-Bolton et al deployed a larger monitoring array in the same residential living room to characterize exposure as a function of the distance from a continuous CO source. Klepeis et al measured real-time CO concentrations at up to 36 points in a residential backyard to consider the proximity effect outdoors near a building. These tracer gas studies provided insight into how different environmental conditions influence the proximity effect; however, they did not account for the characteristics of real smoking or vaping emissions, such as the exhalation of mainstream smoke and the buoyancy of sidestream smoke that can also affect proximity exposure greatly. Studies involving real human smoking or vaping were conducted mostly in prescribed settings. Acevedo-Bolton et al performed controlled experiments inside 2 homes and 16 outdoor locations, using a small group of investigators wearing personal exposure monitors to measure PM2.5 exposure close to prescribed tobacco cigarette smoking. Ott et al used a similar small-group monitoring approach to measure PM2.5 exposure near prescribed tobacco cigarette smoking at 6 outdoor bus stops on California roadways. Zhao et al measured indoor PM2.5 concentrations at 4 different distances from volunteers performing e-cigarette vaping, using a standardized puff frequency indoors in an 80 m3 patient room in a clinical research center. Using a heated mannequin, Martuzevicius et al measured indoor particle exposures at 3 different distances from e-cigarette vaping, adopting the same 30 s puff frequency.

Nguyen et al investigated particle concentrations at personal-space, social-space, public-space distances from non-prescribed vaping activities in California vaping shops. These studies provided valuable data for the levels of exposure close to tobacco smoking or vaping in real-world indoor and outdoor settings. Marijuana is most often smoked in homes . Using a commercial real-time sensor , a recent research study monitored particle number concentrations in ~300 California residences. This study provided the first set of data on particle levels inside real homes with marijuana smoking. However, this large-scale study did not allow spatial measurement of exposure inside a home or accurate mass concentration measurements based on gravimetric calibration. Little is known about the PM2.5 exposure close to a marijuana smoker. There also is virtually no knowledge of how different source types and environments affect the proximity effect. Our first goal was to examine, for the first time, PM2.5 exposure close to a marijuana smoker and how the exposure can be reduced by increasing the distance from the source; we measured real-time PM2.5 concentrations at 1, 2, and 3 m distances from marijuana emissions in a smoker’s home and assessed both the level and frequency of exposure versus distance. Our second goal was to investigate whether choosing a different source type, a different location, or a different environmental setting can reduce the proximity exposure; we tested two common marijuana source types along with their corresponding exhalation patterns in an indoor and an outdoor location under different ventilation and air mixing conditions. Given the collected exposure data, an additional goal of our research was to explore data analysis methods that can potentially be useful for evaluating the recommended physical distance from marijuana sources to minimize involuntary exposure. We performed field research inside a residential property in San Jose, CA . This single-family home has two stories and a private backyard, and the marijuana smoker is the only occupant in this property. Five AM510 SidePakTM monitors were deployed near the indoor chair in the 4.3×3.7×2.4 m family room or the outdoor chair in the backyard where the participant normally smokes or vapes marijuana . Both chairs backed up to a wall, and the outdoor chair had a small table 0.7 m high to its immediate left.

The 5 SidePak monitors were placed radially with 15o angle spacing at an equal distance from the source in each session , measuring PM2.5 concentration every 1 s; they were facing the front of the smoker to account for the worst-case exposure. Three monitors were placed at 1 m height ,grow lights shelves whereas two monitors were at 1.5 m height to consider typical adult breathing heights while sitting and standing, respectively . The actual measured breathing heights of the smoker sitting on the indoor and outdoor chairs were 1.2 m and 1.1 m, respectively. Using these monitoring settings, we performed 35 experiments . For the indoor experiments, 17 were performed with all windows and interior or exterior doors closed in the house – “base case” while 3 involved opening the family-room door and two dining room windows while running the fan of the centralized HVAC system – “alternative case”. For outdoor experiments, 12 were carried out with a fully-opened outdoor umbrella above the smoker – “base case” – while 3 were carried out with this umbrella fully closed – “alternative case”. We hypothesize opening or closing the umbrella would noticeably affect the air mixing and proximity effect close to the source. For the base-case experiments, all 5 monitors were underneath the umbrella when placed at 1 m distance from the smoker. We used the VelociCalc 8386 anemometer to measure and log the indoor and outdoor air velocities near the smoking or vaping locations every 2 s during each experiment. This instrument has a 6-mm diameter sensor probe with a 25 mm long anemometer at its tip, and its minimum detectable air speed is 0.01 m/s. It was not possible to release carbon monoxide or sulfur hexafluoride tracer gas in the participant’s house. As a way to estimate the magnitude of ventilation, we burned matches inside the house while using the Optical Particle Sizer 3330 to measure the particle number concentrations every 1 min. The air change rate was estimated by the log linear regression between concentration of the smallest particle size range and time after the well-mixed condition was reasonably achieved. Given the timescale of the experiments , diffusional and gravitational losses of particles within this size range were expected to be negligible compared with air exchange; this method has been used to estimate ACH in a residence where tracer gas releases were not feasible . These air change rate tests were performed outside the regular smoking or vaping experiments, because they involved particle emissions. We investigated two types of marijuana sources regularly used by the participant: a cigarette-like marijuana joint with 0% CBD and 9.6% THC, and an electronic vaping pen with the “Care by Design” 2:1 cartridge . A standardized smoking or vaping protocol that consisted of 5 puffs over a 5-minute period was used. After inhaling, the participant exhaled at the starting point of every minute ; we defined the 5 min period as the source period. This protocol was intended to enable comparisons between experiments with different source types or source distances based on the same exhalation or emission frequency . Zhao et al and Martuzevicius et al have adopted this approach but with a different frequency for ecigarette vaping. In our study, the participant chose the 1-min time interval for the 5-puff sequence to not exceed his normal habit of smoking and vaping. We did not choose a specific volume and duration for each puff, since we wanted to preserve the behavioral differences embedded in each puff for different source types and to investigate how they may affect the spatial variation of exposure close to a source.

The participant did not permit sensors to be used in contact with his body; therefore, puff topography or spirometry measurement involving sensor mouthpiece breathing was not conducted in this study. As a surrogate approach, we placed the VelociCalc anemometer in front of the smoker during the 5-min source period at 0.1 m horizontal distance from the mouth position to record the “exhalation peak velocity” – the maximum air velocity produced by each exhalation . This approach enabled us to investigate human exhalation via air environment measurement. We discovered the temporal fluctuations of air velocities outdoors were comparable to the magnitudes of exhalation peak velocities. Therefore, we were not able to measure the exhalation peak velocities in the outdoor experiments. The durations of the exhalation were measured by the participant using a stopwatch. A test examining how consistently exhalation peak velocities can be produced and measured by the environmental sensing method is available in the Supplementary Material . PM2.5 Calibration. To ensure consistent measurements between monitors, we conducted a separate quality assurance study in which we placed 17 SidePak monitors inside a car chamber with a smoke source, simultaneously measuring PM2.5 concentrations every 1 min. After the emission stopped and well-mixed condition was reasonably achieved , the exponentially decaying measurements of the SidePak monitors were compared by linear regression with our reference SidePak monitor, giving R 2 > 0.999 for the 5 SidePak monitors used . The SidePak monitors measure PM2.5 concentration based on light scattering properties, which are affected by the particle size and composition. To accurately represent the actual PM2.5 concentration, the calibration factor – the ratio of gravimetrically-to-optically-measuredPM2.5 concentration is needed for each source type . In a previously published paper, we determined the CFs for the two marijuana source types: 0.35 for joint smoking and 0.44 for vaping for the reference SidePak monitor; they were applied along with the inter-monitor slopes to rescale all PM2.5 measurement in this study where i = 1-5. Jiang et al found that CFs for SidePak monitors remained relatively constant over time; for a 16-month period, the average difference was ~3%. The particle zero filter was attached to the inlet of each SidePak monitor immediately before each experiment for zero calibration. The PM2.5 measurements at the 3 different distances were collected from separate experiments, not simultaneously.