Waterpipe Tobaccos, Part I. Composition of Forty Brand-Styles of Contemporary Waterpipe Tobacco Available on the US Market in 2020–2022

Based on references given in Shihadeh et al. (1) and Maziak et al. (2), the predecessors of the shisha products currently on the US market were developed during the 1990's, with the key difference between the new products and the prior products being sufficient glycerol in the composition to prevent combustion of the tobacco and provide the source of the aerosol that was produced when the shisha was heated (3, 4). There were several articles on toxicants in shisha smoke in the following years until the publication by Schubert on relationship between glycerol (VG) and propylene glycol (PG) levels in waterpipe tobaccos and carbonyl levels in smoke when the products were heated (5). However, only two commercial samples from the same manufacturer (Nakhla) were included in the sample set. VG, PG, and nicotine levels in another series of Nakhla products were reported by Wagener et al. (6) in 2019. According to Supplemental Table 1 in Wagener et al. (6), nicotine, propylene glycol, and glycerol by flavor and condition, products listed as having humectants, had VG contents ranging from 15% to 37%, while those products listed as not having humectants, had VG contents ranging from 1.6% to 24%. Data on VG and PG levels in a set of commercial products were reported by Jaccard and coworkers in 2020 (7). However, that report did not identify the brand-styles associated with the analytical data, although it is likely that all brand-styles listed in Jaccard's Table 1 except for T7, T9, and T10 contained sufficient VG and PG to perform in a similar manner to the leading brands of contemporary waterpipe tobaccos available for sale in the US. By contemporary brands, we mean waterpipe tobaccos that do not combust when used as intended in a waterpipe. It is important to note that in US retail markets (local smoke shops as well as internet-based merchants), the terms “waterpipe”, and “waterpipe tobacco”, are not used. A waterpipe is known as a “hookah” or “hookah pipe”. Waterpipe tobacco is known as “shisha”, “hookah tobacco”, or “flavored tobacco”. There is no indication that unflavored shisha is available commercially. However, it can be produced in the laboratory (see “MATERIALS AND METHODS”). The aerosol (smoke) from unflavored shisha is without taste.

Several years ago, we were approached by a representative of several small manufacturers and importers of shisha, and we were asked how they should get information on the chemical composition of their products. We suggested use of an ISO 17025 accredited laboratory that could do the traditional tobacco analytes as well as VG, PG, fructose, glucose, sucrose, and water by the Karl Fischer method. We also suggested that samples of relevant competitive products be sent to the same laboratory too. Furthermore, we obtained samples of the products and the tobaccos used to make them for our own analytical work. In addition, we prepared a flue-cured (FC) surrogate shisha and a dark air-cured (DAC) surrogate shisha. They were sent to the same laboratory for analysis along with samples of the starting tobaccos. The results of the laboratory analyses are shown in Table 1 (see page 153).

The following codes are used in Table 1.

The data in Table 1 show that there were major differences among the waterpipe tobaccos analyzed. These differences will be explained the “RESULTS AND DISCUSSION” section of this report.

As noted earlier, the samples analyzed by the laboratory came from several manufacturers and importers of waterpipe tobaccos. These were contemporary waterpipe tobaccos in that they were all flavored and contained sufficient glycerol or a mixture of glycerol and propylene glycol such that they would be heated, but not combusted during normal operation of a waterpipe. In addition, we asked those organizations to provide samples of their starting tobaccos. We received several samples of flue-cured tobacco and one sample of dark air-cured tobacco. The analytical data we received with the samples are shown below in Table 2.

A typical formula for shisha is 60% glycerol, 20% tobacco, 10% invert sugar syrup, and 10% propylene glycol (PG). This PG simulates the PG and flavor concentrates that are used when making a shisha that has flavors with low transfer rates such as para-tolualdehyde, cis-3-hexen-1-ol, benzaldehyde, para-anisaldehyde, and menthol (8). While glycerol levels in commercial products range from 30 to 70% of product weight, we made our recipe with 60% glycerol (VG).

The experimental procedure for the preparation of the high-VG FC surrogate shisha is as follows:

Glycerol (600 g) and invert sugar syrup (100 g) were mixed in 5-qt Pyrex bowl of KitchenAid KSM155GBCA Mixer and Flat Beater at slowest speed on the mixer.

A similar surrogate product was made with DAC tobacco.

Glycerol (420 g) and invert sugar syrup (70 g) were mixed in 5-qt Pyrex bowl of KitchenAid KSM155GBCA Mixer and Flat Beater at slowest speed on the mixer. Then water (70 g) was added and stirred in.

Unlike the FC surrogate shisha and commercial FC shishas, the DAC surrogate was composed of small particles as is the case with commercial DAC shishas. Also, on standing, there was a noticeable separation between the liquid phase on the bottom of the container and the particles above. This is typical behavior for many commercial DAC shishas. The FC tobacco used in many shisha products is specially grown and processed to have, “high absorption capacity of molasses” (9). Molasses is a shisha industry term that refers to sugar syrups of all types added to the tobacco.

There are several factors to consider when reviewing the data in Table 1. First, the samples are “point-in-time” samples. They may or may not be representative of the products available to consumers. There are no standards for the collection and processing of samples such as there are for cigarettes as given in ISO 8243 (10). Shisha products tend to separate into a tobacco-rich portion on top and a liquid-rich portion on the bottom once the products are packaged. As with sampling, there are no generally accepted procedures for sampling the samples of shisha once they are removed from the packaging. Moreover, there are no standard methods applicable for the sampling and analysis of shisha such as the CORESTA Recommended Methods (CRM) (11). The closest method is CRM 61 (12). The laboratory reported that it used a method similar to that and validated it for the ranges of VG, PG, fructose, glucose, and sucrose found in the shisha samples.

Table 1 shows the results of the analyses performed by the ISO 17025 accredited laboratory. The laboratory selected the analytes that in its view were essential for characterizing waterpipe tobaccos. These analytes included some of the harmful and potentially harmful constituents (HPHC) typically determined on the tobacco-containing products as well glycerol, propylene glycol, fructose, glucose, and water (not oven moisture). There are other analytes in greater concentration than some of those just mentioned, but they were not on the list recommended by the laboratory. These include the higher saccharides such as found in corn syrup, flavor carriers such as ethanol and triacetin, and citrus oils that are both flavors in themselves as well as solvents for other flavors.

Unlike most other tobacco-containing products, the tobacco content of the commercial and surrogate products listed in Table 1 is less than 30% of product weight, and in some cases it is less than 20% of product weight. The purpose of the tobacco is to serve as a carrier for the glycerol that is needed to form the aerosol and to provide a source of nicotine. Glycerol is most always present at more than 30% of product weight, but many products have glycerol contents between 45 and 60%. The main purpose of the glycerol is to provide the aerosol that is the smoke produced by heating the shisha with charcoal briquettes or electric heaters. The reported transfer rate for glycerol is 17% (8) under the conditions specified in ISO 22486:2019 (20) and ISO TS 22487 (21). Thus, a glycerol inclusion rate of 30% should provide sufficient glycerol for a 10 g sample to provide aerosol for smoking under the conditions specified in the aforementioned ISO documents. One manufacturer, M2 in Table 1, added an approximately equal weight of propylene glycol to its loading of glycerol. Propylene glycol has a much lower boiling point than glycerol and a much higher transfer rate than does glycerol (8). This allows the smoker to get early puffs as the shisha is heated to the temperature required for glycerol alone.

Inspection of the data in Table 1 for Manufacturers M1 through M11, show that some manufacturers use a higher proportion of sugar syrups than others and consequently have a lower proportion of glycerol when not used in conjunction with propylene glycol as in the case of M2. Added sugar syrups (i.e., high fructose corn syrup, partially inverted sucrose syrup, corn syrup) have two functions. Generally, 5–10% added sugar syrup is needed to hold the tobacco particles together when the user transfers the product from the package to the bowl of the hookah pipe, and to minimize the loss of tobacco particles during smoking when Egyptian-style bowls are used [Egyptian-style bowls are the commercial equivalent of the waterpipe tobacco holder specified in ISO 22486:2019 (20)]. Added sugars in excess of 5–10% are used to dilute the tobacco so that nicotine concentration is about 0.1%.

Also, unlike most other tobacco products, waterpipe tobaccos require mixing before use, as some of the tobacco particles, even though they are coated with glycerol, sugar syrups, and flavors tend to float on that mixture when allowed to stand for a day or so. However, even if samples are finely ground by processing through a food processor such as the Robot Coupe Blixer 2, there is still separation of the phases. Another factor to consider when reviewing the data in Table 1 is that we have no information on the batch-to-batch variation when two different batches of the same flavor product are analyzed. By different batches, we mean different lots of starting tobacco, different preparation of the base blend, and different flavoring and packaging operation.

The products of the first eleven manufacturers (M1–M11) manufacture or import products based on FC tobacco. Reportedly, the FC tobaccos used in the manufacture of waterpipe tobacco are grown and cured under conditions that keep nicotine levels around 1% and reducing sugar levels in excess of 20% (9). Based on the data in Table 1 and in the data in JACCARD's Table 1 (7) (for all his samples except T7, T9, and T10), most manufacturers of FC products are targeting a nicotine level of 0.1%. Since most manufacturers offer a wide variety of flavors, and some of these flavors involve changing the type and amounts of flavors and flavor carriers, relatively consistent nicotine levels across a manufacturer's product line may be the best indication of manufacturing consistency. Products from manufacturers M1 through M7 exemplified such consistency. Products from M8 showed greater variation not only in nicotine content, but also in levels of glycerol. We do not have any information to explain the compositional variations reported for M8's products. Products from Manufacturers M9 and M10 appear similar to other FC-based products. Products from M11 are unique in that the fructose content is much higher than the glucose and sucrose contents. We have no information to explain the high fructose content found in M11's products.

Manufacturers M12 and M13 specialize in manufacture of waterpipe tobaccos based on dark air-cured tobaccos. There are no detectable sugars in DAC tobaccos. Thus, the sugars found in the one sample from M12 and the two samples from M13 come from added sugars. In the case of M12, the sugar source may have been partially inverted sucrose syrup. The sugar source for the two M13 products cannot be determined based on the available data.

The analytical data for the two surrogate samples compare favorably with the starting tobaccos and ingredients added to those tobaccos. The high tobacco for the DAC surrogate shisha relative to the commercial DAC products may be due to the fact that no citric acid was added to the formulation. Citric acid is believed to be a common additive use in the manufacture of DAC shishas.

The data in this report shows that shisha (waterpipe) tobaccos vary widely in their composition and physical properties. It is likely that such differences in composition will adversely affect any attempts to determine emissions using the instrumentation specified in ISO 22486:2019. Moreover, the data in this report show that the analytical ranges for ingredients added to tobacco during shisha manufacture are outside those that have been validated for existing consensus standard methods. Furthermore, new methods are needed to determine levels of other commonly used ingredients such as corn syrup, triacetin, citric acid, and the citrus oils that are used as both flavors and carriers of other flavors.