The Public Health Gains Had Cigarette Companies Chosen to Sell Very Low Nicotine Cigarettes

Abstract

Introduction

The U.S. Food and Drug Administration (FDA) has proposed lowering the nicotine content of cigarettes to a minimally addictive level to increase smoking cessation and reduce initiation. This study has two aims: (1) to determine when cigarette manufacturers had the technical capability to reduce cigarette nicotine content and (2) to estimate the lost public health benefits of implementing a standard in 1965, 1975, or 1985.

Methods

To determine the technical capability of cigarette companies, we reviewed public patents and internal cigarette company business records using the Truth Tobacco Industry Documents. To evaluate the impact of a very low nicotine content cigarette (VLNC) standard on smoking attributable deaths (SADs) and life-years lost (LYLs), we applied a validated (CISNET) model that uses past smoking data, along with estimates of the potential impact of VLNCs derived from expert elicitation.

Results

Cigarette manufacturers recognized that cigarettes were deadly and addictive before 1964. Manufacturers have had the technical capability to lower cigarette nicotine content for decades. Our model projected that a standard implemented in 1965 could have averted 21 million SADs (54% reduction) and 272 million LYLs (64% reduction) from 1965 to 2064, a standard implemented in 1975 could have averted 18.9 million SADs and 245.4 million LYLs from 1975 to 2074, and a standard implemented in 1985 could have averted 16.3 million SADs and 211.5 million LYLs from 1985 to 2084.

Conclusions

Millions of premature deaths could have been averted if companies had only sold VLNCs decades ago. FDA should act immediately to implement a VLNC standard.

Implications

Prior research has shown that a mandated reduction in the nicotine content of cigarettes could reduce the prevalence of smoking and improve public health. Here we report that cigarette manufacturers have had the ability to voluntarily implement such a standard for decades. We use a well-validated model to demonstrate that millions of smoking attributable deaths and life-years lost would have been averted if the industry had implemented such a standard.

Introduction

We have known for decades that nicotine is the primary addictive component of tobacco.¹ Recently, the U.S. Food and Drug Administration (FDA) announced intentions to regulate the nicotine content of cigarettes for the purposes of improving public health.² Indeed, a mandated reduction in the nicotine content of cigarettes could reduce the prevalence of smoking by both increasing smoking cessation³ and decreasing the progression of smoking from experimental use to regular use and dependence.⁴ Clinical trials assessing the impact of such a policy have shown that switching smokers to very low nicotine content cigarettes (VLNCs) compared with smokers continuing to smoke normal nicotine content cigarettes (NNCs) reduced the number of cigarettes smoked per day, lowered levels of dependence, and increased quit attempts.^5–8 Future public health benefits were demonstrated in a recent simulation study, which indicated that a mandated reduction of nicotine in 2016 would prevent 8.5 million deaths by the year 2100.⁹

The internal business records of cigarette manufacturers have shown that the companies knew in the 1950s and early 1960s that cigarettes were the cause of cancer and that nicotine addiction was the main motivation for persistent smoking leading to premature deaths from smoking addiction.^10–14 The commercial feasibility of manufacturing and marketing a VLNC was also demonstrated in the early 1990s, when Philip Morris briefly test marketed denicotinized cigarettes under various brand names (i.e., Next, Benson & Hedges De-Nic, Merit De-Nic).¹⁵ Philip Morris withdrew their VLNCs from test markets claiming that they had not been commercially successful. However, a careful review of their own marketing data reveals that market share was up to 0.4% in some test markets, which would be equivalent or greater than many cigarette brands that have been on the market for years.¹⁶ A few years later, a genetically engineered VLNC called Quest 3 was sold by Vector Tobacco between 2004 and 2009. These examples demonstrate the technical feasibility of creating and marketing VLNCs and raise doubts about statements made by manufacturers that it would require decades to introduce VLNCs consistent with the FDA’s proposed very low nicotine standard for combustible tobacco products.

The purpose of this article is to examine when cigarette manufacturers first had the technical capability to reduce the nicotine of cigarettes they produced and to quantify the public health impact of the decision of cigarette manufacturers to continue to produce and market highly addictive cigarettes rather than switch production to VLNCs. A Very Low Nicotine Model considers the impact on reducing smoking and the associated reduction in premature smoking attributable deaths (SADs) and life-years lost (LYLs) that would have occurred as a result of implementing a nicotine standard as early as 1965. We consider public health benefits if only VLNC were marketed in 1965, 1975, or 1985.

Methods

Historical Analysis of Cigarette Companies

To assess when cigarette manufacturers had the technical capability to lower nicotine in the tobacco used in cigarettes, we conducted a review of public patents on nicotine removal in tobacco and internal cigarette company business records accessed from the Truth Tobacco Industry Documents website housed at University of California San Francisco.¹⁷ Examples of search terms used for identifying patents and industry documents included the following: nicotine, nicotine yield, reduction of nicotine, reduced nicotine, very low nicotine, extraction of nicotine, extraction of alkaloid, remove nicotine, alkaloid-reduced nicotine, control of nicotine, reconstituted tobacco, genetic engineering, denicotinize, chemical and thermal methods, supercritical extraction, nonaddictive tobacco, modifying nicotine, low nicotine blend, methyl ethyl ketone, selective extraction, and Project 0302. A snowballing procedure¹⁸ was utilized to follow-up on patents, documents, names, and projects numbers that provided leads for further searching.

Modeling the Impact of Industry Behavior

Our analysis estimating the public health benefits of a VLNC standard begins with a Status Quo Scenario based on traditional NNC use, which is compared with a Reduced Nicotine Scenario characterized by replacing NNCs with VLNCs. The Status Quo Scenario is based on actual smoking rates in the initial year, and projections into future years are based on smoking initiation, cessation, and mortality rates. Under the Very Low Nicotine Scenario, fewer never smokers initiate smoking and more smokers quit in reaction to VLNCs replacing NNCs. Projected outcomes include changes in smoking prevalence, SADs, and LYLs. Our analyses in each of the scenarios are applied over a 100-year period.

Status Quo (Normal Nicotine Levels) Scenario

The model is initialized in 1964 with separate prevalence rates classified by never, current, and former cigarette smokers. Prevalence data, by age and sex in 1964, were obtained from the National Health Interview Survey (NHIS). Current smokers were measured as those who had smoked 100 cigarettes in their lifetime and currently smokes cigarettes every day or some days.

To project smoking rates forward from 1964, we apply age-, sex-, and year-specific initiation, cessation, and mortality probabilities. Estimates of initiation and cessation probabilities by age, sex, and birth cohort were developed by applying an age–period–cohort statistical model to data from the 1965–2018 NHIS while correcting for bias due to higher mortality among current and former smokers over age 40.^19–23 Initiation is based on respondents’ stated ages of initiation. Cessation is measured as the percent quit from smoking for at least 2 years to approximate cessation net of relapse, that is, assuming those who quit less than 2 years are offset by those who relapse after 2 years. The initiation and cessation probabilities were validated by comparing the projections over the period 1965 through 2018 against smoking prevalence rates.^22,23 A percentage of never, current, and former smokers die each year based on age-, sex-, and year-specific mortality probabilities distinguished by smoking status.^20,21

In developing the model, a fixed population based on Census Bureau projections²⁴ is used to estimate the number of current and former smokers. SADs are estimated by multiplying the number of current and former smokers by their excess risks, that is, for current (former) smokers calculated as the difference between the current (former) smoker and never smoker mortality rates. LYLs are estimated by multiplying the premature deaths at each age by the expected years of life remaining of a never smoker at that age.

The Impact of VLNCs

Because VLNCs have not yet replaced NNCs, direct estimates produced by a voluntary very low nicotine standard for combusted cigarettes are not directly observable. However, randomized controlled trials that have examined the effects of VLNCs generally obtain potent effects on cessation.^3,25–27 Studies also indicate potentially large effects on mitigating youth and young adult initiation⁴ and young adult smoking.²⁸

Based a panel of experts, specific effect sizes for replacement of NNCs with VLNCs were recently developed by Apelberg et al.⁹ Applying an expert elicitation (EE) technique, experts were asked to estimate a range of probabilistic estimates of the impact on initiation, cessation, and product switching in response to a very low nicotine standard on cigarettes, with other tobacco products, including smokeless tobacco and electronic (e)-cigarettes, not covered by the standard. For the 50th percentile with ranges of 25th–75th percentile, the panel estimated that 19% (11%–30%) of male and 21% (12%–30%) of female smokers would quit in the first year and 13.5% (8.4%–23.8%) of male and 15% (9.4%–26.3%) of female smokers would quit in the following years, and smoking initiation would be reduced by 50% (25%–65%) for both sexes in the first year and by 30% (20%–57.5%) in future years. Conducted in early 2017, the experts predicted that 40% (30%–55%) of male and 35% (25%–52.5%) of female smokers would switch completely to noncombustible nicotine sources (including e-cigarettes).

We rely primarily on the EE⁹ for the impacts of an industry standard for VLNCs. Besides using the midpoint estimate from the EE as the most plausible parameters, we used the 25th and 75th percentile estimates from the EE lower and upper parameters. Although the bulk of smoking initiation dating back to 1965 took place by age 34, significant initiation occurred after age 34, particularly among women. Consequently, we increased the initiation multipliers for those ages. Cessation rates are based on those who quit all tobacco use, but the effects of current smokers who switch to the less harmful noncombusted products are incorporated into our upper bound cessation estimates. We distinguished cessation multipliers for the first year that a nicotine standard is in place from later years. Because EE participants were told that the cessation rate in the absence of VLNCs was 3.7%, we developed cessation as multiples of that rate, but because age-specific cessation rates were generally much higher (as much as 10% from 1965 through 2000) after age 64, we reduced the cessation multiplier at ages above 64. Because former smoker death rates reflect the distribution of former smokers by years quit and those who quit before age 40 will have risks at later ages close to those of never smokers, smokers who quit before age 40 due to VLNCs are assumed to have mortality rates equal to those of never smokers. From age 40, those who quit become former smokers.

Our initiation and cessation midpoint, lower bound and upper bound multipliers are shown in Supplementary Table 1. Compared with midpoint estimates, lower bound impacts yield less reduction in initiation and less increase in cessation and upper bound estimates yield greater reduction in initiation and greater increase in cessation. In terms of the initiation in the Status Quo Scenario, midpoint male and female smoking initiation rates under the Very Low Nicotine Scenario for all ages and all years are reduced by 50% for the midpoint estimates for both sexes (35% for males and 30% for females lower bound and 75% for both sexes upper bound) in all years. For cessation rates, those age 65 and above were distinguished from those below age 65 due to their substantially higher cessation rates, earlier years were distinguished from later years due to the potentially higher impact in the first years. Cessation rates are increased by 500% for midpoint estimates (200% lower bound and 900% upper bound) for ages less than 65 and 200% for midpoint estimates (100% lower bound and 400% upper bound) for elder ages in the initial year and increased by 200% (100% lower bound and 400% upper bound) for ages less than 65 and 100% (50% lower bound and 150% upper bound) for elder ages in following years.

Public Health Impacts

We conducted three sets of analyses with start dates when NNCs are replaced by VLNCs: 1965, 1975, and 1985. Upon applying the reduced initiation and increased cessation VLNC parameters, the methods described above in the Status Quo section for determining SADs and LYLs are then applied to current and former smoking prevalence.

The public health impacts are gauged in terms of the relative difference in current and former smoking prevalence (i.e., (Rate (VLNCs) − Rate (SQ))/Rate (SQ), where VLNCs is from the Very Low Nicotine and SQ is from the Status Quo Scenarios) and the absolute difference in SADs and LYL. The impact of VLNCs on smoking prevalence and deaths is projected over a 100-year horizon, in order to incorporate the effects on mortality of those at early ages when initiation takes place. The cumulative public health impact is determined by summing the averted deaths and LYLs between the Status Quo and Very Low Nicotine Scenarios from 1965 to 2064 with VLNCs beginning in 1965, from 1975 to 2074 with VLNCs beginning in 1975, and from 1985 to 2084 with VLNCs beginning in 1985.

Results

Feasibility of Reducing the Nicotine Content of Tobacco Filler Used in Making Cigarettes

Table 1 summarizes patents and industry documents on nicotine removal and the design objective of ensuring a minimum amount of nicotine in tobacco filler to ensure that cigarettes could create and sustain nicotine dependence.^29–41 We found patents and internal company documents dating back to the 1920s and 1930s, which describe methods to extract nicotine from tobacco.^29–31

Table 1.

The Role of Nicotine in Cigarette Design: Evidence of Patents and Corporate Documents

Date	Source	Document title or quotea
July 2, 1929	Patent28	Patent no. 1719291: Removing nicotine from tobacco. https://pdfpiw.uspto.gov/.piw?Docid=1719291&idkey=NONE&homeurl=http%3A%252F%252Fpatft.uspto.gov%252Fnetahtml%252FPTO%252Fpatimg.htm
May 7, 1935	Patent29	Patent no. 2000855: Method of denicotinizing tobacco. https://pdfpiw.uspto.gov/.piw?Docid=2000855&idkey=NONE&homeurl=http%3A%252F%252Fpatft.uspto.gov%252Fnetahtml%252FPTO%252Fpatimg.htm
December 9, 1935	ATC30	It is quite possible to “denicotinize” a cigarette by chemical and thermal methods. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/sxwv0024
November 2, 1959	RJR31	The physiological requirements of the smoker with respect to nicotine can be met by the application of the optimum amount of nicotine to the extracted tobacco. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/fxkp0034
July 17, 1963	B&W32	Moreover, nicotine is addictive. We are, then, in the business of selling nicotine, an addictive drug effective in the release of stress mechanisms. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/rhxp0042
September 18, 1963	B&W33	It may be well to remind you, however, that we have a research program in progress to obtain, by genetic means, any level of nicotine desired. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/jglw0200
February 1, 1965	PM34	Determine minimum nicotine required to keep human smoker “hooked.” https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/qynn0226
May 24, 1971	RJR35	Habituating level of nicotine (how low can we go?). https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/tjvk0191
April 14, 1972	RJR36	Research activities need to more precisely define the minimum amount of nicotine required for “satisfaction” in terms of dose levels, dose frequency, dosage form and the like. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/stdb0184
July 22, 1977	Lorillard37	The level of nicotine in the smoke required to produce the desired results is an unknown factor; however, we have estimated it to be in the neighborhood of 0.4–0.5 mg per cigarette. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/gxbv0035
June 30, 1978	PM38	For quite some time, we have been interested in finding out how smokers would react to cigarettes with normal levels of tar and very low levels of nicotine….The tobacco used to make these cigarettes was treated for 25 min with steam and ammonia to reduce the total alkaloids from 1.56% to 0.07%. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/srdv0184
February 13, 1980	Lorillard39	Determine the minimum level of nicotine that will allow continued smoking. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/kpmv0035
January 18, 1982	PM40	…a threshold [for nicotine effects] exists somewhere between 0.1 and 0.3 mgs of nicotine. https://www.industrydocumentslibrary.ucsf.edu/tobacco/docs/tnbx0108

^aLinks to U.S. patents and/or documents from the Truth Tobacco Industry Documents collection.

Case I: VLNCs Beginning in 1965 (Impact for 1965–2064)

As shown in Table 2, male smoking prevalence in the Status Quo Scenario declines from 56.9% in 1964 to 34.8% in 1990 and 6.3% in 2064, whereas male former smoking prevalence increases from 16.1% in 1964 to 25.3% in 1990 and 25.5% in 2015 and then declined to 20.6% in 2064. Female smoking prevalence was 35.4% in 1964 declining to 30.3% in 1990 and 5.8% in 2064, whereas former smoking prevalence increases from 5.3% in 1964 to 21.3% in 2015 and then declines to 13.8% in 2064. From 1965 to 2015, cumulative SADs are estimated as 20.4 million (14.4 million for males; 6.0 million females), and LYLs are 254.1 million (178.6 million males; 75.5 million females), similar to an earlier study.¹⁹ From 1965 to 2064, SADs are 38.4 million (26.5 million males, 11.9 million females) and LYLs are 423.8 million (293.7 million males, 130.1 million females).

Table 2.

Population Impacts on Smoking Prevalence, Smoking Attributable Deaths, and Life-Years Lost in the Status Quo Scenario and Very Low Nicotine Scenario (Midpoint, Lower Bound, and Upper Bound) When Very Low Nicotine Content Cigarettes Begin in 1965

		Prevalence				Cumulative impacta
		1964	1990	2015	2064	1965–1990	1965–2015	1965–2064
Male
Status quo	Current	56.9%	34.8%	20.0%	6.3%	—	—	—
	Former	16.1%	25.3%	25.5%	20.6%	—	—	—
	SADs	250 701	284 725	295 371	172 499	7 191 782	14 370 583	26 505 181
	LYLs	2 956 834	3 406 323	3 892 702	1 337 864	85 813 480	178 617 477	293 682 931
VLNC: midpoint	Current	56.9%	12.1%	4.0%	1.3%	−65.3%	−80.0%	−79.5%
	Former	16.1%	24.9%	13.8%	4.9%	−1.7%	−45.8%	−76.1%
	Av SADs	0	109 115	169 830	133 762	2 143 512	5 461 853	13 922 577
	LYGs	0	1 813 044	2 919 507	1 092 477	31 828 245	93 283 758	184 122 616
VLNC: lower bound	Current	56.9%	19.4%	8.1%	2.4%	−44.2%	−59.7%	−62.1%
	Former	16.1%	26.0%	18.7%	9.0%	2.5%	−26.7%	−56.1%
	Av SADs	0	76 664	123 647	100 269	1 323 666	3 730 136	9 989 982
	LYGs	0	1 240 005	2 179 537	856 810	19 426 216	63 799 750	134 256 734
VLNC: upper bound	Current	56.9%	4.4%	1.0%	0.4%	−87.4%	−94.8%	−93.6%
	Former	16.1%	21.5%	6.9%	1.5%	−15.3%	−72.8%	−92.6%
	Av SADs	0	129 515	209 341	161 211	2 889 890	6 885 155	17 382 498
	LYGs	0	2 233 913	3 454 269	1 266 465	43 941 649	117 868 590	224 382 828
Female
Status quo	Current	35.4%	30.3%	16.4%	5.8%	—	—	—
	Former	5.3%	16.3%	21.3%	13.8%	—	—	—
	SADs	36 062	143 063	140 210	69 488	2 466 932	6 019 333	11 874 669
	LYLs	535 523	1 693 556	1 802 867	506 462	32 660 960	75 509 106	130 100 508
VLNC: midpoint	Current	35.4%	10.9%	3.0%	1.1%	−64.0%	−81.8%	−80.6%
	Former	5.3%	17.6%	12.4%	3.7%	8.1%	−41.8%	−73.4%
	Av SADs	0	65 996	86 991	54 470	709 421	2 743 448	6 979 170
	LYGs	0	930 881	1 446 923	433 375	11 093 005	42 058 134	87 770 714
VLNC: lower bound	Current	35.4%	18.2%	6.7%	2.3%	−40.0%	−59.0%	−60.3%
	Former	5.3%	17.6%	16.8%	7.0%	8.3%	−20.9%	−48.9%
	Av SADs	0	42 163	65 896	39 875	417 804	1 877 048	5 032 362
	LYGs	0	595 637	1 095 339	343 862	6 498 068	28 666 477	64 466 439
VLNC: upper bound	Current	35.4%	3.9%	0.7%	0.3%	−87.0%	−95.7%	−94.2%
	Former	5.3%	16.2%	6.8%	1.1%	−0.7%	−68.2%	−91.9%
	Av SADs	0	84 414	99 593	65 101	1 008 200	3 377 404	8 432 576
	LYGs	0	1 203 894	1 629 609	486 213	16 111 287	52 286 157	103 633 965

VLNC = very low nicotine cigarette; SADs = smoking attributable deaths; Av SADs = averted SADs; LYL = life-years lost; LYG = life-years gained.

^aCumulative impact includes (1) the reduction in current and former smoking prevalence in the Very low Nicotine Scenario relative to the initial value in the Status Quo; (2) the cumulative deaths (SADs) and life-years lost (LYLs) in the Status Quo Scenario; and (3) the cumulative averted deaths (Av SADs) and life-years gained (LYGs) are those the Very Low Nicotine Scenario subtracted from those to the Status Quo.

Using midpoint parameters for the Very Low Nicotine Scenario, male smoking prevalence declined to 12.1% in 1990, 4.0% in 2015, and 1.3% in 2064 (an 80% relative reduction). Female smoking prevalence declined to 10.9% in 1990, 3.0% in 2015, and 1.1% in 2064 (an 81% relative reduction). Compared with the Status Quo Scenario, 8.2 million SADs (5.5 million males, 2.7 million females) and 135.3 million LYLs are averted by 2015, and 20.9 million SADs and 271.9 million LYLs are averted by 2064. Using lower bound parameters, male smoking falls to 2.4% in 2064, whereas female smoking falls to 2.3% in 2064. By 2064, 15.0 million SADs and 198.7 million LYL are averted. Using upper bound parameters, male smoking declines to 4.4% in 1990 and 0.4% in 2064, whereas female smoking declines to 3.9% in 1990, and 0.3% in 2064. By 2064, 25.8 million SADs and 328.0 million LYLs are averted.

Case II: VLNCs Beginning in 1975 (Impact for 1975–2074)

As shown in Table 3, Status Quo Scenario smoking prevalence was 47.6% for males and 33.8% for females in 1974, falling to 6.1% and 5.6%, respectively, by 2074 as former smoking prevalence rose and then fell. From 1975 to 2074, we estimate 36.7 million SADs and 393.5 million LYLs. Using VLNC midpoint parameters, male smoking prevalence declines to 1.3% in 2074, whereas female smoking prevalence declines to 1.1% in 2074. Compared with the Status Quo Scenario, 18.9 million SADs and 245.4 million LYLs are averted by 2074. Using lower bound parameters, smoking prevalence declines less rapidly than the midpoint case, with 13.5 million SADs and 178.6 million LYLs averted by 2074. Using upper bound parameters, smoking prevalence sees more rapid declines, with 23.3 million SADs and 296.3 million LYLs averted by 2074.

Table 3.

Population Impacts on Smoking Prevalence, Smoking Attributable Deaths, and Life-Years Lost in the Status Quo Scenario and Very Low Nicotine Scenario (Midpoint, Lower Bound, and Upper Bound) When Very Low Nicotine Content Cigarettes Begin in 1975

		Prevalence				Cumulative impact*
		1974	2000	2025	2074	1975–2000	1975–2025	1975–2074
Male
Status quo	Current	47.6%	29.3%	13.8%	6.1%	—	—	—
	Former	21.2%	25.0%	25.6%	19.6%	—	—	—
	SADs	274 864	281 293	284 616	156 062	7 332 447	14 489 786	25 211 934
	LYLs	3 339 400	3 535 630	3 243 106	1 204 818	87 864 270	179 793 553	271 258 561
VLNC: midpoint	Current	47.6%	9.7%	2.2%	1.3%	−67.0%	−84.3%	−79.3%
	Former	21.2%	24.1%	13.0%	4.5%	−3.3%	−49.4%	−76.9%
	Av SADs	0	101 200	149 527	124 898	2 115 806	5 261 098	12 480 675
	LYGs	0	1 911 148	2 339 206	994 517	34 268 128	93 778 040	164 868 096
VLNC: lower bound	Current	47.6%	16.0%	4.7%	2.3%	−45.3%	−65.6%	−61.9%
	Former	21.2%	25.2%	17.9%	8.4%	4.6%	−30.0%	−57.2%
	Av SADs	0	70 992	109 052	95 390	1 323 583	3 598 022	8 909 514
	LYGs	0	1 298 618	1 781 861	785 386	21 139 041	64 224 911	119 688 177
VLNC: upper bound	Current	47.6%	3.3%	0.6%	0.4%	−88.6%	−96.0%	−93.5%
	Former	21.2%	20.9%	6.8%	1.4%	−17.0%	−73.5%	−92.8%
	Av SADs	0	120 423	185 413	147 024	2 787 972	6 581 924	15 570 559
	LYGs	0	2 365 561	2 746 420	1 143 089	46 250 375	117 511 969	200 889 870
Female
Status quo	Current	33.8%	24.6%	11.4%	5.6%	—	—	—
	Former	9.2%	18.9%	21.2%	12.7%	—	—	—
	SADs	70 394	139 034	135 130	56 951	3 105 530	6 465 107	11 481 055
	LYLs	1 024 543	1 609 776	1 604 887	444 489	38 446 371	80 404 880	122 203 403
VLNC: midpoint	Current	33.8%	7.8%	1.7%	1.1%	−57.2%	−85.1%	−80.6%
	Former	9.2%	19.3%	11.7%	3.3%	2.0%	−44.9%	−74.0%
	Av SADs	0	73 357	79 315	46 227	1 087 240	2 926 236	6 403 687
	LYGs	0	1 035 823	1 260 621	385 071	15 661 959	46 279 689	80 581 631
VLNC: lower bound	Current	33.8%	13.8%	4.1%	2.2%	−36.6%	−64.1%	−60.2%
	Former	9.2%	19.8%	16.1%	6.4%	2.8%	−24.0%	−49.7%
	Av SADs	0	49 333	60 613	34 452	659 837	2 014 527	4 591 722
	LYGs	0	691 872	974 932	307 438	9 448 135	31 819 726	58 924 936
VLNC: upper bound	Current	33.8%	2.5%	0.4%	0.3%	−75.2%	−96.4%	−94.2%
	Former	9.2%	17.2%	6.6%	1.0%	−12.9%	−68.9%	−92.0%
	Av SADs	0	88 383	91 137	53 924	1 468 790	3 569 528	7 755 786
	LYGs	0	1 269 727	1 408 200	428 257	21 622 275	56 790 671	95 397 053

VLNC = very low nicotine cigarette; SADs = smoking attributable deaths; Av SADs = averted SADs; LYL = life-years lost; LYG = life-years gained.

^aCumulative impact includes (1) the reduction in current and former smoking prevalence in the Very Low Nicotine Scenario relative to the initial value in the Status Quo; (2) the cumulative deaths (SADs) and life-years lost (LYLs) in the Status Quo Scenario; and (3) the cumulative averted deaths (Av SADs) and life-years gained (LYGs) are those the Very Low Nicotine Scenario subtracted from those to the Status Quo.

Case III: VLNCs Beginning in 1985 (Impact for 1985–2084)

As shown in Table 4, Status Quo Scenario smoking prevalence was 38.3% for males and 30.2% for females in 1985 and falling to 6.0% and 5.6%, respectively, by 2084 as former smoking prevalence rose and then fell. From 1985 to 2084, we estimate 34.6 SADs and 360.0 million LYLs. Using midpoint parameters for VLNCs, male smoking prevalence declines to 1.3% and female smoking prevalence declines 1.1% by 2084, leading to total of 16.3 million SADs and 211.5 million LYLs averted by 2084. Using lower bound parameters, 11.6 million SADs and 152.8 million LYLs are averted. Using upper bound parameters, 20.1 million SADs and 257.2 million LYLs are averted.

Table 4.

Population Impacts on Smoking Prevalence, Smoking Attributable Deaths, and Life-Years Lost in the Status Quo Scenario and Very Low Nicotine Scenario (Midpoint, Lower Bound, and Upper Bound) When Very Low Nicotine Content Cigarettes Begin in 1985

		Prevalence				Cumulative impacta
		1984	2010	2035	2084	1985–2010	1985–2035	1985–2084
Male
Status quo	Current	38.3%	23.0%	9.8%	6.0%	—	—	—
	Former	23.8%	25.3%	25.0%	19.1%	—	—	—
	SADs	293 713	277 917	266 517	140 838	7 402 872	14 446 092	23 926 684
	LYLs	3 452 870	3 758 528	2 494 757	1 106 055	92 848 239	175 020 600	249 947 585
VLNC: midpoint	Current	38.3%	6.5%	1.6%	1.3%	−71.8%	−84.2%	−79.2%
	Former	23.8%	24.6%	11.9%	4.4%	−2.7%	−52.6%	−77.1%
	Av SADs	0	97 462	112 980	114 188	1 969 861	4 657 324	10 788 561
	LYGs	0	1 987 480	1 635 166	917 122	35 504 840	84 454 901	141 932 779
VLNC: lower bound	Current	38.3%	11.7%	3.2%	2.3%	−49.1%	−67.1%	−61.7%
	Former	23.8%	25.5%	16.5%	8.1%	0.9%	−33.9%	−57.6%
	Av SADs	0	67 008	82 036	88 312	1 207 009	3 159 122	7 671 983
	LYGs	0	1 335 622	1 263 731	727 749	21 483 723	57 391 419	102 322 668
VLNC: upper bound	Current	38.3%	1.9%	0.4%	0.4%	−91.6%	−95.4%	−93.5%
	Former	23.8%	21.8%	6.8%	1.4%	−13.9%	−72.9%	−92.8%
	Av SADs	0	117 818	143 389	132 991	2 648 680	5 884 081	13 487 304
	LYGs	0	2 471 537	1 936 857	1 050 105	48 845 449	106 894 149	174 224 056
Female
Status quo	Current	30.2%	18.5%	8.3%	5.6%	—	—	—
	Former	13.3%	20.3%	19.8%	12.1%	—	—	—
	SADs	110 276	125 285	128 213	50 040	3 287 700	6 505 976	10 662 178
	LYLs	1 451 209	1 607 498	1 238 318	403 852	40 277 255	78 485 526	110 013 751
VLNC: midpoint	Current	30.2%	5.0%	1.3%	1.1%	−72.8%	−84.5%	−80.6%
	Former	13.3%	20.4%	10.8%	3.2%	0.7%	−45.3%	−73.9%
	Av SADs	0	61 430	63 601	41 086	1 183 954	2 776 873	5 466 989
	LYGs	0	1 030 903	892 943	351 161	17 836 252	44 403 732	69 547 665
VLNC: lower bound	Current	30.2%	9.6%	2.9%	2.2%	−47.9%	−64.6%	−60.2%
	Former	13.3%	21.0%	14.7%	6.1%	3.8%	−25.7%	−49.6%
	Av SADs	0	42 202	48 822	30 934	719 291	1 908 987	3 895 300
	LYGs	0	695 776	703 727	281 352	10 779 356	30 463 873	50 445 154
VLNC: upper bound	Current	30.2%	1.4%	0.4%	0.3%	−92.3%	−95.7%	−94.2%
	Former	13.3%	18.4%	6.8%	1.0%	−9.2%	−65.8%	−92.0%
	Av SADs	0	72 645	73 969	47 512	1 587 175	3 393 696	6 649 224
	LYGs	0	1 250 409	1 000 678	389 493	24 351 522	54 605 585	82 971 737

VLNC = very low nicotine cigarette, SADs = smoking attributable deaths, Av SADs = averted SADs; LYL = life-years lost; LYG = life-years gained.

^aCumulative impact includes (1) the reduction in current and former smoking prevalence in the Very Low Nicotine Scenario relative to the initial value in the Status Quo; (2) the cumulative deaths (SADs) and life-years lost (LYLs) in the Status Quo Scenario; and (3) the cumulative averted deaths (Av SADs) and life-years gained (LYGs) are those the Very Low Nicotine Scenario subtracted from those to the Status Quo.

Discussion

A prior modeling study found that adopting a government-imposed cigarette nicotine reduction standard today would avert 8.5 million fewer deaths and 33.1 million life years lost by 2100.⁹ Based on our review of public patents on nicotine removal from tobacco and company business records, it is apparent that commercially feasible designs for VLNCs have existed for decades, long before FDA was given regulatory authority to regulate tobacco and proposed a rule to set a standard for the amount of nicotine allowed in tobacco used in cigarettes.^{15,16,29–41} Indeed, cigarette companies had the opportunity to market VLNCs right after the 1964 Surgeon General Report made clear that cigarettes being manufacturers at the time were deadly.⁴² Applying this same model a well-validated simulation model, we estimated that 20.9 (15.0–25.8) million SADs and 271.9 (198.7–328.0) million LYLs would have been averted over 100 years had cigarette companies chosen to only manufacture and sell VLNCs starting in 1965. This represents a 54% reduction in SADs and a 64% reduction in LYLs using midpoint estimates. If the very low nicotine standard were implemented in 1975, 18.9 (13.3–23.3) million SADs and 245.4 (178.6–296.3) million LYLs would have been averted over 100 years. Delaying to 1985, the number of SADs averted falls to 16.3 (11.6–20.1) million and the numbers of LYLs averted falls to 211.5 (152.8–257.2) million.

Our results provide evidence for the potential public health benefits associated with a very low nicotine standard. In retrospect, it is apparent that the industry did not convert production to very low nicotine, less addictive cigarettes, in order to protect the profits which they recognized were tied to selling addictive products. Instead, it appears that they chose to provide a cigarette that is designed to show reduced nicotine and tar yields via smoking machines but engineered in ways that allowed smokers to easily compensate to obtain enough nicotine, thereby allowing cigarettes to remain addictive and difficult to quit.⁴³ It is also likely that the high levels of industry concentration and strong barriers to entry into the cigarette industry⁴⁴ protected the existing firms from competition from manufacturers producing very low nicotine products as well as from firms selling lower risk alternative nicotine delivery products.⁴⁵ In the absence of competition, the industry was in a position to collude. Indeed, the cigarette industry was involved in a conspiracy to conceal the truth about the harms of smoking,^10–14 which provided them an excuse to maintain the status quo. Had any one of the companies revealed the truth about the health harms and addictiveness of cigarettes, manufacturers may have been forced to modify their cigarettes in order to mitigate the health impacts.

Rather than maintaining the status quo, established cigarette firms could have used their market power to support a very low nicotine standard and worked with the government to set and enact such a standard.^46–48 Allying with government, public education programs⁴⁹ could have been used to encourage smokers to switch to VLNCs or quit altogether. At the same time, a very low nicotine standard may have protected the cigarette industry from lawsuits and public condemnation. Indeed, the public health harms from not producing VLNCs as documented in this study should expose cigarette manufacturers to increased liability.

Our study is subject to limitations. The effect sizes for VLNCs are directly based on an EE conducted by the FDA.⁹ This process includes supporting evidence-based primarily clinical trials, where NNCs were available and noncompliance may have occurred.⁵⁰ However, many of the studies on cessation^3,25–27 are based on self-reported responses from smokers on how they would react to a VLNC policy. In addition, human studies have not explicitly considered the impacts of VLNCs on smoking initiation for ethical reasons. Our estimates of the effects of VLNC standard on smoking initiation and transitions to regular smoking and dependence is limited because the studies were either conducted with animals or in laboratory settings. Nevertheless, to incorporate uncertainty, we conducted sensitivity analyses in which we incorporated the uncertainty as found in the EE and incorporated additional uncertainty in gauging the effects of VLNCs on initiation. In addition, we used VLNC parameters based on recent estimates. The effects of a VLNC standard may differ if the standard had been implemented earlier, in light of the changes in the tobacco/nicotine product landscape and the current availability of other alternatives, such as e-cigarettes. However, the EE attempted to distinguish switching to alternative delivery products and our results are based on primarily those who quit all nicotine use.⁹

A further limitation is that the status quo results are based on actual smoking rates over the period from 1964 to 2018 and projections based on that data. We did not explicitly incorporate how the policies implemented and their impact on smoking rates may have differed if a voluntary nicotine standard had been adopted. We also did not consider that smokers may substitute other forms of tobacco, such as smokeless tobacco and cigars, in response to a voluntary nicotine standard. In that case, fewer tobacco SADs and LYLs would be averted. However, if the risk of using these other products is less, there would still be health gains. In addition, possibly less harmful and addictive versions of these other products may have come onto the market. Finally, we did not incorporate the potential role of compensation,⁵¹ such as inhaling deeper or smoking cigarettes closer to the butt of the cigarette, because evidence indicates that any compensation with VLNCs is temporary.^51–54 Although all of these limitations merit consideration, we did conduct extensive sensitivity analysis and found that even more pessimistic assumptions about the impact of yield considerable public health gains.

In summary, we have shown the immense public health gains that would have been possible had cigarette manufacturers self-regulated and adopted a voluntary very low nicotine standard. The huge toll on SADs could have been reduced by 54% and LYLs by 64% if a voluntary standard was adopted in 1965. The gains would have been a smaller proportion, though still substantial if a VLNC standard had been adopted in 1975 or 1985. These results demonstrate the lives lost as a result of choices by the cigarette manufacturers and the need for rapid action to mitigate continued public health detriment.

Supplementary Material

A Contributorship Form detailing each author’s specific involvement with this content, as well as any supplementary data, are available online at https://academic.oup.com/ntr.

Funding

This project was funded through a P01 grant (P01CA200512) from the National Cancer Institute, Naitional Insitutes of Health. D.L. and R.M. received funding from the National Cancer Institute through a TCORS grant U54CA229974 and a CISNET Lung Cancer grant U01CA199284. Salary support was provided by the National Institute on Drug Abuse for T.S. (K01DA047433) and B.W.H. (K23DA041616).

Declarations of Interests

K.M.C. has served as an expert witness in litigation against the cigarette industry. All other authors have no conflicts of interest to declare.