Measures for Assessing Subjective Effects of Potential Reduced-Exposure Products

Potential reduced-exposure products (PREP) developed by the tobacco industry may reduce toxicant exposure, and it is hypothesized that this can reduce health risks. Types of PREPs include oral noncombustible products such as snuff or snus (e.g., Marlboro and Camel Snus) that have varying levels of toxicant content, compressed tobacco lozenges (e.g., Stonewall, Ariva, or Camel Orbs), and other products such as tobacco strips or sticks. Other kinds of PREPs are modified cigarettes or novel cigarette delivery devices. Modified cigarettes may have reduced toxicants in the tobacco product itself (e.g., tobacco specific nitrosamines) or a special filter that reduces volatile toxicants. Examples of these modified cigarettes are Advance, Omni, and Marlboro UltraSmooth which are no longer on the market. Novel cigarette delivery devices include devices that heat rather than burn tobacco so reduced amounts of combustion products and their associated toxins are emitted (e.g., Accord, Eclipse), or electronic “cigarettes” that deliver nicotine in a vapor (1). These types of products are marketed to be used in lieu of or in addition to smoking cigarettes.

However, whether these products lead to reduced health risk to the individual or reduced harm to the population compared with the use of conventional tobacco products is uncertain. Moreover, if people believe that PREPs are a safer alternative to conventional tobacco products, PREPs could cause harm by discouraging smokers from quitting or persuading nonsmokers to use PREPs (2).

Scientists both inside and outside the tobacco industry are evaluating PREPs to determine the potential health effects associated with their use, using laboratory and clinical designs (3, 4). The goal of laboratory studies is to examine the acute effects such as short-term physiologic responses, blood nicotine and carbon monoxide levels, and subjective responses. Some laboratory studies are also conducted to determine the abuse liability or abuse potential of the product, that is, the likelihood of uptake and persistent use of the product (5). In these trials, both subjective and behavioral responses to a product are assessed. Subjective responses such as product liking and the good or bad effects of the product are often used in abuse liability studies because of their high reliability and face validity. These measures can be used during the laboratory administration of the product or they can be used retrospectively to understand the complete experience of using the product and to determine the likelihood of using the product after being abstinent from the product for a period of time. Sensory stimuli associated with positive or negative responses to smoking that may also contribute to the abuse liability of a product can be observed by use of product effect scales with items such as “stimulated,” “dizzy,” or “heart pounding” (5). Behavioral responses can include whether the individual chooses one product or another and/or the extent the individual self-administers a product. Typically, in laboratory studies, products are compared against other PREPs, placebo, and/or usual brand products. Clinical trials, on the other hand, are conducted in the natural environment or in a residential facility and similarly compare the effects of a PREP against other PREPs, placebos, or usual brands over a longer term or with a greater extent of use. In these clinical trials, the outcome measures include the same types of measures used in laboratory studies, but may also focus on the pattern of product use, the extent of toxicant exposure, and biomarkers that measure toxicity or biological effects. A critical review and understanding of the strengths and limitations of measures used to assess product effects is important given the newly granted Food and Drug Administration authority to regulate tobacco products, including modified-risk products.

An important and common assessment measure that is employed across different types of studies examining PREPs is the participant's subjective responses to the PREP. These measures broadly encompass self-reported expressions of the participant's experience using the product. The types of measures used to assess PREPs include product evaluation scales, sensory evaluation scales, drug-liking and drug effect ratings, and withdrawal scales described in Table 1. Investigators have developed their own scales or have used well-known measures such as the Minnesota Nicotine Withdrawal Scale (MNWS), the Cigarette Evaluation Questionnaire, or the Direct Effects Scale. However, no consensus measure for the subjective effects of PREPs has been developed, complicating cross-study comparisons.

Further, no systematic review of the various scales used by research groups has been conducted. This type of review is critical in providing sound measures to help determine the abuse liability of a product, a major area of PREP evaluation. Evaluation of both positive (e.g., liking the product, satisfaction from the product) and negative reinforcing effects (e.g., reducing negative affect, withdrawal symptoms, and craving) that may sustain the product's use is an important part of assessing the potential harm of a product. Products that are high in toxicant exposure and in abuse potential are likely to produce significant harm to the individual and the population. On the other hand, products with low toxicant exposure and low abuse potential are likely to lead to less harm. Measuring the subject's response to the effects of the product is just one area of assessment, and it complements assessment of the consumer's perception of the product as a result of marketing, labeling, and patterns of use in social networks. Reviews of methods and measures for evaluating tobacco consumer response and their application to assessing PREPs (4), and reviews of existing surveillance systems that enable the identification and tracking of tobacco products including PREPs (6) have recently been conducted.

The purpose of the present paper is to: (a) identify scales that are used to evaluate PREPs and describe the items on the scales; (b) evaluate whether the scales and items are sensitive in testing PREPs (that is, whether they can detect differences across products including placebo); and (c) recommend scales to investigators to assess PREPs or identify items that may be critical for the development and validation of new scales.

The criteria for evaluating the sensitivity of these subjective effects measures and items in detecting differences across products include: (a) whether the scale distinguishes placebo or no-tobacco conditions from the PREP; (b) whether the scale or item shows that PREPs are rated as good or less good than the usual brand of cigarettes; and (c) whether the scale or item distinguishes between PREP products. Where PREP studies and results exist, the relationship between the scales or items with behavioral measures will be described. This review is not intended to comprehensively examine the validity and reliability of the scales.

PubMed was searched on April 14, 2008 and on June 1, 2009, with the search limited to human studies published in English. The following search terms were used: reduced exposure products and tobacco; specific PREP product names (e.g., Eclipse, Accord, Advance cigarettes, Ariva, Snus, Stonewall, Quest, Next cigarettes, Omni); denicotinized cigarettes and tobacco; light cigarettes and tobacco; ultralight cigarettes and tobacco; low tar and tobacco; low yield and tobacco; electrically heated cigarettes; and smokeless and tobacco. Studies were included in the review if they assessed a PREP and used at least one subjective measure. The data were compiled to examine: (a) the extent the scale was sensitive in measuring subjective responses to the PREP, and (b) whether the scale was directly associated with other outcome measures or behavior. Three major categories of subjective effects measures were identified: craving, withdrawal symptoms, and product responses.

Several different scales were used to measure withdrawal symptoms and craving. Many studies used the MNWS (7) or a modified version of this scale; refs. 8-26). Items were rated as follows: 0, not present; 1, slight; 2, mild; 3, moderate; or 4, severe, or as visual analogue scales where items were presented above a horizontal line with anchors on the left (“not at all” or “none”) and the right (“extremely” or “severe”).

Many other studies used the Questionnaire of Smoking Urges (QSU; ref. 27) or a short version of this scale (28) as a measure of urges to smoke (11-16, 18-20, 22, 23, 26, 29, 30). The QSU includes 32 items (or 10 items on the short version) that are rated on a 7-point scale ranging from 0 (strongly disagree) to 6 (strongly agree). The scale yields two factors: intention to smoke (factor 1) and anticipation of relief from withdrawal (factor 2).

The Shiffman-Jarvik Withdrawal Scale (31) or an abbreviated version was used in four studies (19, 29, 32, 33). This scale has 25 items presented in a 7-point scale, with 1 indicating “very definitely” and 7 indicating “very definitely not.” The items are divided into the five following subscales based on factor analysis: craving, psychological discomfort, physical symptoms, stimulation/sedation, and appetite.

A Desire to Smoke visual analogue scale that was derived from Schuh and Stitzer (34) included four or five items and was used in four studies (19, 21, 29, 30). Visual analog scales were presented as a 100-point horizontal line, anchored on the left side with “not at all” and on the right side with “very much.” Fagerström, Hughes, Rasmussen, and Callas (35) measured withdrawal symptoms on a 5-point scale derived from the American Psychiatric Association, and two studies used a nonreferenced scale (21, 36).

There were two aims for using withdrawal symptom scales in PREP studies: (a) to measure the extent the product reduces withdrawal symptoms from cigarettes; and (b) to determine withdrawal symptoms from the product. To date, no studies have examined withdrawal symptoms from a PREP.

A few laboratory studies reported the effect of withdrawal symptoms and craving using a modified cigarette such as Advance (14), smokeless tobacco (ST) PREPs (Ariva, Marlboro Snus, Camel Snus) and a modified cigarette (Quest) (18), or ST PREPs such as Stonewall and General Snus (20) compared with a placebo condition. Two of three of these studies found that compared with PREPs, the placebo condition showed fewer decreases in withdrawal symptoms or craving ratings (14, 18). The laboratory component of another study, however (20), reported decreased withdrawal symptoms for a nontobacco placebo smokeless product, usual brand, and loose moist snuff (General snus), but not in a ST PREP condition (Stonewall) for the item “craving a dip.” Additionally, the QSU factor 1 score decreased significantly after the first use of the nontobacco placebo smokeless product and a loose moist snuff and after the fourth use in all conditions.

A few clinical studies compared PREPs with no smoking or no ST and found that ratings of withdrawal and craving in no-tobacco conditions were at times higher than those in the PREP conditions, including two studies that used modified cigarettes such as Advance (12) or denicotinized cigarettes (29), one study that used modified and heated cigarettes such as Advance and Eclipse (11), and one study that used ST PREPs such as Stonewall and General Snus (20).

Several laboratory studies that compared PREPs with the subjects' usual brand reported that there were fewer decreases in ratings of withdrawal symptoms and craving in the PREP conditions compared with the usual brand, including one study that used heated cigarettes such as Accord (15) and two studies that used modified and heated cigarettes such as denicotinized cigarettes, Accord, and Eclipse (13), or denicotinized cigarettes and Accord (16). Alternatively, a different laboratory study that used denicotinized cigarettes compared with the subjects' usual brand or a light brand cigarette found that there were no significant brand effects or brand by time interactions for withdrawal symptoms (21).

A clinical study that used reduced-nicotine cigarettes (9) reported some significant differences in ratings of withdrawal (e.g., irritability and increased eating) between reduced-nicotine cigarettes and usual brand, whereas three other clinical studies that used modified cigarettes such as Omni (22) or Advance (12) or heated and modified cigarettes such as Advance and Eclipse (11) reported that, in general, there were no differences between PREPs and the usual brand.

Several laboratory studies indicated a decrease in withdrawal symptoms and craving ratings, and no significant differences in the extent of reduction between denicotinized and nicotinized cigarettes (17, 19, 23, 26, 30).

A few clinical studies also indicated no significant differences in ratings of withdrawal symptoms or craving between PREPs or as the dose of the PREP changed, including one study that used increasing amounts of heated cigarettes (Accord) with the option of using usual brand cigarettes (37), one study that used heated cigarettes (Eclipse) as against a nicotine inhaler (36), and one study that used ST PREPs (Exalt and Ariva) compared with medicinal nicotine lozenges (25).

However, some laboratory studies reported greater decreases in withdrawal symptoms as the dose of the PREP increased or among products that had higher doses of nicotine levels, including two studies that used ST PREPs (10, 24). For example, in one laboratory study, reduction in withdrawal symptoms was observed in a ST PREP study, in which smokers were administered Ariva tablets (one, two, and three tablets in ascending order), at 90-minute intervals (10). Mean withdrawal scores decreased significantly following three Ariva tablets. In another laboratory study among ST users, single doses of ST PREPs (Ariva, Stonewall, Revel), medicinal nicotine (Commit lozenge), and moist snuff (Copenhagen) were administered (24). Withdrawal symptoms were lower during Copenhagen (high-nicotine product) use than with Revel (lower-nicotine product) use. Craving was lower in Copenhagen compared with the other four products and was lower in Stonewall, Ariva, and Commit compared with Revel. There were no other significant differences between products for withdrawal symptoms or craving.

Additionally, a clinical study using Eclipse compared with a nicotine inhaler also found some differences between products with regard to withdrawal symptoms, but there were no differences in craving between products (35).

Generally, laboratory studies that used more than one withdrawal symptom and craving scale showed similar overall results across scales (13-16, 18, 19, 26, 30). Likewise, for most clinical studies that used more than one withdrawal symptom and craving scale, the measures showed comparable results (11, 12, 20, 22).

Some studies related the results of the withdrawal symptom scales to other outcomes. For example, in a laboratory trial of ST PREPs, medicinal nicotine, and Copenhagen among ST users, Kotlyar and colleagues (24) found a significant negative correlation between nicotine area under the concentration time curve (AUC) and craving score, but not between nicotine AUC and withdrawal symptoms. In another laboratory study, Dallery and colleagues (19) used a within-subject design in which subjects participated in one session each of rapid smoking (up to 9 cigarettes with puffs every 6 seconds) and normal-paced smoking with nicotinized and denicotinized cigarettes and then were given a choice to smoke. They reported that craving ratings were significantly higher when subjects chose to smoke versus when they declined a chance to smoke.

There are many studies that measure withdrawal symptoms and craving, with the MNWS or an adapted version of the scale used most frequently in PREP studies. Most laboratory and clinical studies that compared PREPs with a placebo or no-tobacco condition showed differences between PREPs and a placebo or no-tobacco condition. Withdrawal symptoms and craving ratings showed fewer decreases in the placebo or no-tobacco condition. On the other hand, in one of the laboratory studies comparing ST PREPs with a placebo, scores lessened for the item “craving a dip” in the placebo condition, but not for a ST PREP condition, and showed decreases in QSU factor 1 scores in the placebo and General snus conditions after the first use compared with decreases in a ST PREP and usual brand condition after the fourth use (20). Moreover, when PREPs were compared with the subjects' usual brand, studies reported that withdrawal symptoms and craving ratings were generally lower for the subjects' usual brand or that there was no difference in ratings. Whether scales distinguish between PREPs depended on the product used or how much product was used. For example, there were no significant differences between denicotinized and nicotinized cigarettes regarding the extent of withdrawal symptoms and craving reduction in several laboratory studies or between PREPs in a few clinical studies. The lack of differences in withdrawal and craving suppression between denicotinized and nicotinized cigarettes points to the important role of sensory factors associated with these abstinence effects (38). Some studies, however, reported differences in withdrawal symptoms and craving relief across different PREPs and across different doses of nicotine. For example, two ST PREP studies showed greater withdrawal symptoms or craving relief as the dose of Ariva increased (10) or for the product containing the most nicotine (24). For most laboratory and clinical studies that used more than one scale, there were no meaningful differences in ratings of withdrawal symptoms or craving between scales.

Few studies have examined the relationship between ratings on withdrawal and craving to other measures. Of the two existing studies, a relationship was observed between withdrawal symptoms or craving ratings and nicotine AUC levels (24) or smoking behavior (19).

In summary, withdrawal symptoms and craving assessments, with studies predominantly using the MNWS, have been found to be in the expected direction in some cases but not in others. Withdrawal and craving symptoms are more severe in placebo or no-smoking conditions than with PREPs. On the other hand, usual cigarettes do not always tend to reduce symptoms more than PREPs, and PREPs with higher nicotine doses do not always reduce symptoms to a greater extent than PREPs with lower nicotine doses. These equivocal results may reflect the lack of sensitivity of the scales, no detectable differences between PREPs and usual brand cigarettes and among PREPs, or the fact that withdrawal is strongly associated with the sensory and behavioral aspects of smoking. Few studies have been conducted on the relationship between self-reported withdrawal and craving and behavioral or physiologic response.

In contrast to the measures of withdrawal and craving, no single scale to measure product effects has been used predominately in studies examining the use of modified cigarettes, heated cigarettes, or ST PREPs. Many studies used scales that asked a combination of questions regarding the liking of the product, including items such as “desire for the product” or “satisfaction”; sensory and physical effects of the product such as “nausea,” “heart race,” or “dizziness”; and product evaluation such as “tobacco strength” or “smoothness” (9, 11, 18, 19, 20, 29, 39). Other studies used scales that examined only the liking of the product and the product evaluation (17, 21, 26, 32, 40-42) or the liking of the product and the sensory and physical effects of the product (8, 33, 43). A few studies used scales that only examined the sensory and physical effects of the product (10, 18, 19), one study used a scale that examined the liking of the product (25), and three studies used a scale that evaluated the product and the sensory and physical effects of the product (33, 40, 44).

There were three aims for using scales to measure the effects of the product in PREP studies: (a) to determine the extent to which the product is rewarding; (b) to determine the sensory and physical effects of the product; and (c) to assess the characteristics of the product. Items that showed a positive finding are listed by study and product type for each category in Table 2. Items with nonsignificant results are not presented because the items that showed a positive finding may be ones that should be included in future scales and validated in future studies. However, a lack of significant findings may not necessarily indicate that the item should not be included in future studies.

In both a laboratory study that used modified (denicotinized) cigarettes (17) and a laboratory study that used ST PREPs (Ariva, Camel Snus, Marlboro Snus) and a modified (Quest) cigarette (18), the PREP was rated higher than the placebo or no-tobacco condition on measures of product liking.

In several laboratory studies subjects preferred their usual brand compared with one or more PREPs on measures of product liking, including studies that used modified cigarettes such as reduced-nicotine-content cigarettes (40) or Marlboro UtlraSmooth cigarettes (with carbon filter; ref. 43), and a study using ST PREPs such as Stonewall and General Snus (20). Moreover, in the clinical component of this study using ST PREPs (Stonewall, General snus), subjects preferred their usual brand over General snus, but not Stonewall, for the item “do you like the way this tastes?” (20). Similarly, subjects rated usual brand higher than PREPS on measures of product liking in a clinical study that used modified reduced-nicotine-content cigarettes (9) and in a clinical study that used heated (Eclipse) and modified cigarettes (Advance; ref. 11).

In four laboratory studies, subjects rated nicotinized cigarettes higher than denicotinized cigarettes in terms of satisfaction, pleasantness, drug effect, or taste (17, 19, 32, 41). In other laboratory studies, differences were also observed on measures of liking, satisfaction, and psychological reward. For example, highly ventilated cigarettes were rated higher than cigarettes with low nicotine content (0.02 mg nicotine) on these items (33). In another laboratory study that examined the effects of two nonnicotine cigarettes with differing amounts of tar in which half of the subjects received the low-tar cigarette and half received the high-tar cigarette, the low-tar cigarette was rated higher on a measure of good drug effect compared with the high-tar cigarette (39). Another study observed that the direction of these ratings may depend on the amount of nicotine in the product (26). Higher ratings of “bad effects” and “taste” were observed with full-tar conventional cigarettes compared with full-tar denicotinized cigarettes, and higher satisfaction with full-tar denicotinized products. On the other hand, higher ratings of bad effects and lower ratings of taste were observed with reduced-tar denicotinized cigarettes compared with reduced-tar conventional cigarettes.

In a study that used three Quest cigarettes with different nicotine levels, the 0.6-mg nicotine cigarette was rated as more satisfying than the 0.3- and 0.05-mg nicotine cigarettes (42). Similarly, in a laboratory study among ST users comparing ST PREPs (Ariva, Stonewall, Revel), Copenhagen, and Commit lozenge (all products with different nicotine yields), a significant difference between products was seen during use of Copenhagen (the product containing the highest nicotine) on measures of feeling good effects from the product, satisfaction, liking, and desirability, but not between the other products (24).

Two clinical studies were conducted among smokers to examine Exalt ST or Ariva versus medicinal nicotine lozenges on measures of liking, desirability, and effectiveness. As to having more good effects, subjects rated medicinal nicotine lozenges higher than Exalt (25). Ariva was rated as more desirable than medicinal nicotine lozenges, but there was no significant difference found for liking and effectiveness.

In a ST PREP laboratory study (44) there were differences in ratings between PREPs and a nontobacco condition on measures of sensory and physical effects. This study conducted among ST users was designed to assess nicotine levels and physiologic and subjective effects of ST products, including Copenhagen, Skoal Long Cut Cherry, Skoal Original Wintergreen, and Skoal Bandits or nontobacco mint snuff. Subjects rated Copenhagen highest in increased salivation, nausea, heart racing, head rush, anxiety, and feeling alert (44). Scores on these items were lower for Skoal Long Cut Cherry and Skoal Original Wintergreen, and lowest for mint snuff and Skoal Bandits, with no significant differences between Skoal Bandits, which achieved the lowest nicotine level across the tobacco products, and mint snuff.

In a laboratory study that used research cigarettes with varying levels of nicotine (1 mg, 2 mg, 4 mg, 8 mg, and 12 mg) compared with the subjects' usual brand, subjects gave similar ratings for cigarettes with 8 and 12 mg nicotine and the usual brand of cigarettes for feeling stimulated and feeling heart beating fast, but gave lower ratings for these items when smoking cigarettes with 1 mg nicotine (40). In another laboratory study among smokers involving seven conditions (Ariva, Marlboro Snus, Camel Snus, Commit nicotine lozenge, usual brand, Quest, and a sham cigarette), subjects reported significant increases compared with baseline for the items “calm you down,” “help you concentrate,” “reduce hunger,” “dizzy,” and “more awake” for the usual brand, with not many or no higher ratings in the other noncombustible products. The item “salivation” increased significantly compared with baseline in the noncombustible products only and not in the usual brand condition (18). In a different laboratory study, subjects rated Marlboro UltraSmooth as having a lesser effect compared with their usual brand, but the same effect as Marlboro Ultra Lights (43).

A laboratory component of this ST PREP study among ST users comparing Stonewall and General snus to the subjects' usual brand or placebo found higher ratings for the subjects' usual brand compared with other conditions for the following items: heart race, head rush, relaxed, and alert (20). In the clinical component of this ST PREP study, in which subjects used smokeless tobacco products ad libitum (usual brand, Stonewall, General snus) or used no ST, General snus was scored highest for nausea, followed by Stonewall and then the usual brand. For salivation, Stonewall was rated highest, then the usual brand, with General snus the lowest (20). Another clinical study with four conditions (Advance, Eclipse, usual brand, or no smoking) determined that for the item “do the cigarettes make you sick?” usual brand received the lowest score, followed by Advance, with Eclipse scoring significantly higher than Advance. There was also a significant condition effect for items that measured calming down, concentration, awake, and reduce hunger. Scores were typically highest for the subjects' usual brand and lowest for Eclipse (11).

A few laboratory studies that used modified cigarettes (19, 33, 39, 40) and two ST PREP studies (10, 24) determined that there were differences between products on measures of sensory and physical effects. In one laboratory study, highly ventilated filter cigarettes were compared with low-nicotine-content cigarettes and were found to be rated higher in enjoyment of respiratory tract sensations whereas the low-nicotine-content cigarettes were rated as having greater intensity of respiratory tract sensations depending on the area of the respiratory tract affected (33). In a laboratory study of two nonnicotine cigarettes with differing amounts of tar, a low-tar cigarette was rated higher than a high-tar cigarette on the item “stimulated” (39). In another laboratory study that used nicotinized and denicotinized cigarettes, rapid smoking of nicotinized cigarettes significantly increased all ratings of nicotine effect items. Rapid smoking of denicotinized cigarettes significantly increased ratings of nausea, dizziness, light-headedness, burning throat, racing heart, and headache (19). For nausea and dizziness, the smoking pace with either cigarette influenced the rating, with the nicotinized cigarettes scoring highest. A different laboratory study used research cigarettes with varying levels of nicotine (1 mg, 2 mg, 4 mg, 8 mg, and 12 mg) compared with the subjects' usual brand, and subjects reported significantly higher ratings of feeling high and light-headed or dizzy after smoking high-nicotine cigarettes (40).

In a ST PREP laboratory study that used Ariva, Stonewall, Revel, Commit, and Copenhagen among ST users, there were no significant differences on some items (alert, relaxed, head rush, tremor in hands, arms or face; light-headed/dizzy; drowsy, energetic or jittery; ref. 24). On other items, however, Copenhagen was rated significantly higher than Commit or Revel for fast/pounding heart and significantly higher than Commit for feeling high. In another ST PREP laboratory study designed to determine the effects of Ariva on cigarette smokers, subjects reported feeling more nauseous as the dose of Ariva increased (10). Other items such as feeling dizzy, confused, lightheaded, and nervous were scored higher (although less consistently) after taking Ariva compared with deprivation prior to taking Ariva.

A laboratory study reported differences between ST PREPs and a no-tobacco condition with regard to product characteristics. Copenhagen, Skoal Long Cut Cherry, Skoal Original Wintergreen, and Skoal Bandits or nontobacco mint snuff were compared among ST users. Copenhagen scored highest for the item “product strength” and lowest for “how well packed” (44).

In several laboratory studies that compared modified cigarettes (21, 40), heated cigarettes (45), ST PREPs (20), or ST PREPs and a modified cigarette (18) with the subjects' usual brand, subjects generally rated their usual brand most positively in terms of product characteristics. For example, one laboratory study examined denicotinized cigarettes, light brand cigarettes, and usual brand, and found that subjects rated their usual brand higher on items that measured “cigarette strength” and “more smoke than air” compared with the other conditions (21). The subjects' usual brand was also rated more favorably compared with two heated cigarettes that were considered weaker, more difficult to draw from, and less smooth (45). In a different laboratory study, subjects rated low-nicotine-content cigarettes milder, smoother, less flavorful, of poorer quality, and having less nicotine than their usual brand (40). In another laboratory study among smokers involving seven conditions (Ariva, Marlboro Snus, Camel Snus, Commit nicotine lozenge, usual brand, Quest, and a sham cigarette), subjects reported significant increases in the usual brand condition compared with baseline for the item “does the product taste like your own brand?” (18).

In the laboratory component of this ST PREP study using usual brand, Stonewall, General snus, or placebo among ST users, subjects rated their usual brand higher than the other conditions on the items “tobacco pack” and “tobacco strength.” The item assessing the amount swallowed was rated significantly higher for Stonewall than the usual brand (20). In the clinical component of this ST PREP study during 5-day conditions (usual brand, Stonewall, General snus, and no ST), significant conditions by day interactions were observed for items measuring tobacco strength and tobacco pack (20). Subjects rated these items similarly each day for usual brand, and lower most days for Stonewall and General snus. For the item “amount swallowed” Stonewall was rated highest, followed by usual brand, then General snus.

Another clinic study using four conditions (Advance, Eclipse, usual brand, or no smoking) reported significant condition effects for the following items: “Do the cigarettes taste like your usual brand of cigarette?” “Do the cigarettes feel like your usual brand of cigarette?” and “Do the cigarettes feel as mild as your usual brand of cigarette?” (11). Scores were typically highest for the subjects' usual brand and lowest for Eclipse. Similarly, in a different clinical study, subjects rated reduced-nicotine cigarettes as less strong, less flavorful, and of generally lower quality than their usual brand (9).

A laboratory study that measured ST PREPs (24) and three laboratory studies that used modified cigarettes (39, 41, 42) found differences in subject ratings of PREPs with regard to product characteristics. For example, compared with ST PREPs, Copenhagen was rated as strongest among ST users (24). In a laboratory study of two nonnicotine cigarettes with differing amounts of tar, subjects gave higher ratings for harshness, heaviness, and intensity of flavor to the high-tar versus the low-tar cigarette (39). In another laboratory study, subjects rated nicotinized cigarettes higher in smoothness compared with denicotinized cigarettes (41). Alternatively, in different laboratory studies, there was no significant difference on measures of product characteristics in three studies that used denicotinized versus nicotinized cigarettes (19, 26, 32).

Many different scales were used in PREP studies to evaluate the effects of the product, although there was no scale that was used predominately. The Direct Effect Scale was used in three studies (11, 18, 20), but each used some items that were different. A couple of studies also used the Cigarette Evaluation Questionnaire (8, 33) or the Direct Effects of Nicotine scale (10, 18). Scales used in studies that asked questions about product liking, sensory and physical effects, and product evaluation might provide the most information (see refs. 9, 11, 18-20, 29, 39). In general, the scales or items measuring these effects seem to be sensitive in showing differences between the placebo or no-tobacco conditions and a PREP. Furthermore, the usual brand is more highly rated or rated similarly to PREPs (greater liking, stimulatory or central nervous effects, and greater strength or quality). The scales or items also distinguish between PREP products, with products having higher nicotine content rated more highly than products with lower nicotine content, with the exception of scales measuring product evaluations where the results are equivocal.

Items that showed a significant difference between PREPs compared with a placebo, usual brand, or other PREPs were examined in terms of the three aims for using the scales in PREP studies. It may be worthwhile including these items in future scale development to determine their validity and reliability. For the aim of determining the extent to which the product is rewarding, the items that were most often rated significantly different with regard to PREPs, usual brand, or placebo conditions were as follows: satisfying, less satisfying, taste good/like taste, liking, like/good drug effect, dislike/bad drug effect, desirability, and pleasant. Other items that were rated significantly different in at least one study were dislike taste, like cigarette, dislike cigarette, not as good, unpleasant, psychological reward, and effectiveness.

For the aim of determining the sensory and physical effects of the product the following items were most often significantly different among PREPs, usual brand, and placebo conditions: nausea, dizziness, heart racing, head rush, calm you down, alert, help you concentrate, more awake, reduce hunger, salivation, and feel high. Other items that were significantly different in at least one study were heart pounding/fast, anxious, relax, stimulated, make you sick, enjoyment of respiratory tract sensations, intensity of respiratory tract sensations, less intensity of drug effect, and light-headed.

The last aim of PREP effect scales is to assess the characteristics of the product. The following items were most often significantly different among PREPs, usual brand, and placebo conditions: strength, less flavorful, smoother, poorer/lower quality, and cigarettes taste like your usual brand of cigarette. Other items that were rated as significantly different in at least one study were the following: intensity of flavor, tobacco pack, how well packed, less strong, weaker, milder, more smoke than air, more difficult to draw from, less smooth, harshness, less nicotine, amount swallowed, cigarettes feel like your usual brand, cigarettes feel as mild as your usual brand, and heaviness. The items selected in a study will also vary according to the products being tested.

Scales to evaluate responses to tobacco products are still at the early stages of development, and no concerted attempt has been made to validate the existing scales or to develop scales that may be better predictors of product use behavior. A couple of studies related the results of product effect scales to other outcomes. For example, in a laboratory study with smokers of Marlboro Lights that compared a carbon filter PREP (Marlboro UltraSmooth) with Marlboro Ultra Lights in 48-hour conditions, low subjective ratings for Marlboro UltraSmooth seemed to be associated with low levels of exposure (43). In another laboratory study conducted among ST users designed to assess nicotine levels and the physiologic and subjective effects of ST products, including Copenhagen, Skoal Long Cut Cherry, Skoal Original Wintergreen, and Skoal Bandits or nontobacco mint snuff, increases in the rating of product strength showed an association with plasma concentration during the first 10 to 15 minutes after product administration (44).

In summary, several items measuring rewarding effects, sensory and physical effects, and product characteristics are able to distinguish the effects from different products and also produce results in the expected direction, that is, PREPs are rated more highly than placebos, usual brands are typically rated more highly than PREPs, and high-nicotine-content products produce more central nervous system effects than low-nicotine-content products.

There are many scales used to test PREPs that focus on withdrawal symptoms and craving and the effects of the PREP. Scales used in PREP studies seem to measure withdrawal symptoms and craving effectively. The scales are able to detect decreases in withdrawal symptoms and craving, differences between PREP product and usual brand cigarettes or between PREP products based on nicotine content of the product. This review also showed that when studies used more than one scale, the results revealed a concordance among the scales with a similar pattern of results. Therefore, each of the scales used in PREP studies may be an effective measure of withdrawal symptoms and craving. In the future, PREP studies should consider using standardized measures to allow comparisons across studies and products. To date, the most widely used scale has been the MNWS or its modified version. The second most used scale, the QSU, yields two factors, intention to smoke (factor 1) and anticipation of relief from withdrawal (factor 2), which provides a more thorough examination of urges to smoke. Therefore, at a minimum, these two scales may be the ones that should be included in a battery of assessments for PREPs. Future studies should be directed towards examining if the extent of withdrawal relief from a product is associated with how much the product will be used.

To date, no studies have examined the effects of abstaining from PREPs after long-term use. This area of research is important to provide further validation of scales. For example, it would be hypothesized that withdrawal from oral, low-nicotine products would produce less withdrawal than from a modified cigarette product with high levels of nicotine.

Another key area for evaluating PREPs is the subjective response to a product, which also plays a key role in the evaluation of abuse liability of a tobacco product (5). As described in the article on the assessment of abuse liability of tobacco products, subjective responses to tobacco products are measured across various research paradigms, including acute dose-effects comparisons, drug-discrimination, self-administration, forced-drug choice, and clinical trials, and serve as core tools to determine whether the product is likely to be used, the extent the use of the product is likely to be maintained and to lead to addiction, and the difficulty consumers may experience when trying to stop using the product.

When considering the three primary areas of tobacco product evaluation, most laboratory and clinical studies reported that when comparing PREPs with placebo or no tobacco condition or with the subjects' usual brand cigarettes, the results are in the anticipated direction; generally PREPs are rated more highly than placebo and usual brand more highly than PREPs. Studies also found that for some items there were differences detected between PREPs, but for other items there were no differences between PREPs. Discrepancies in ratings were observed depending on smoking pace (19) or product dose (10).

Unlike the withdrawal symptoms and craving scales, no particular scale was used predominately to measure the effects of PREPs. However, assessing the three primary areas of tobacco product evaluation is critical. Development and validation of a product scale that measures these three areas should be a high priority research area. The scales should have core elements that are comparable across products as well as assessments of unique product-specific attributes that would not necessarily generalize across products, but are important pieces of information to ascertain, particularly for novel designs. Validation of these scales can include examining how effective the scales are in predicting the amount of product self-administered, product choice, and discrimination across products with different nicotine yields or sensory properties, sustained use of the product, and potential dependence and abstinence effects from the product. Future studies can also examine product marketing, messaging, and labeling, price, social norms, and attitudes towards the product effect responses to these scales.

The determination of the validity of these scales, which is the relationship between subjective responses to actual use, is a critical area of research. In addition, the assessment of subjective responses to products in nonsmokers is also important. How we can introduce cigarette PREPs to nonsmokers is a quandary. Another critical area of research is determining how a population of nontobacco users may respond to PREP tobacco products, which would provide information on the potential uptake of these products among this population. This type of research can be conducted with occasional tobacco users or with products that are not highly likely to produce addiction such as nicotinized modified-risk cigarettes. For example, studies have been conducted with medicinal nicotine products with nonsmokers (46) or nicotine nasal spray (47, 48) in subjects naïve to tobacco use.

With the passing of the Family Smoking Prevention and Tobacco Control Act that gives the Food and Drug Administration jurisdiction over evaluation of products, it becomes vitally important that tools are available to effectively evaluate PREPs as well as other tobacco products. Assessing the abuse potential of a product involves determining the subjective responses to these products that are related to behavioral responses. Through these tools we will be able to assure the protection of public health.

Dorothy Hatsukami received grant funding from Nabi Biopharmaceuticals to evaluate the nicotine vaccine. No other potential conflicts of interest were disclosed.

We thank Dr. Mark Parascandola for his valuable comments on the manuscript.