Research Associate at the Department of Pharmacy, University of Parma

Sara Nicoli

Smoking-related illness is a major public health problem, and there is great impetus to develop effective therapies to help stop smoking and maintain abstinence.

  Two types of pharmacological therapies have been approved by FDA for smoking cessation:

  nicotine replacement therapy (NRT), which substitutes nicotine from cigarettes with safer nicotine-containing formulations; and bupropion administration,

  which is an atypical antidepressant that can significantly increase abstinence rate. 

NRT is designed to help smokers quit by reducing withdrawal symptoms.

  All commercially available forms of NRT are effective as part of a strategy to promote smoking cessation,

  even though the quit rate is far from 100% (ranging from 14% to 25% after 12 months). Bupropion is a safe and effective non-nicotine alternative first-line treatment.

Bupropion is a weak inhibitor of dopamine and noradrenaline re-uptake, and has been shown to antagonize nicotinic acetylcholine receptor function.

  When administered with intensive behavioural support, bupropion is as effective as NRT and also reduces the weight gain that often occurs after a smoker quits. 

Because of the limited quit rate with single therapy, a combined therapy of nicotine and bupropion has also been tested in several clinical trials.

  Treatment with oral sustained-release bupropion combined with a nicotine patch resulted in a higher rate of long-term smoking cessation compared with a placebo, and also with a single treatment of either nicotine or bupropion alone. While nicotine is currently administered in different formulations (i.e., transdermal patches, inhalers, chewing-gums, tablets), bupropion is available in tablet form only. The development of a transdermal dosage form containing both actives could be of great interest to the study of smoking cessation. Additionally, bupropion transdermal administration could avoid the risk of seizures associated with peak plasma concentration.

The aim of this work was to evaluate the transdermal penetration characteristics of the two drugs — alone and in combination — and to prepare and test a transdermal system that could deliver both nicotine and bupropion through the skin. Rabbit ear skin was used as the barrier in permeation experiments as it can be regarded as a reasonable model for human skin 

Nicotine was purchased from Merck (Darmstaadt, Germany) and bupropion HCl from Sigma (St. Louis, USA). Polyvinyl alcohol (PVA) (MW=83400; degree of hydrolysis 86–89 mol%) was obtained from Nippon Ghosei (Osaka, Japan) and Polyvynil pyrrolidone K90 (PVP) was received from BASF (Ludwigshafen, Germany). For high performance liquid chromatography (HPLC) analysis, acetonitrile (HPLC grade) (Carlo Erba; Milan, Italy) and distilled water were used. All other chemicals used were of analytical grade. 

 Rabbit ear skin was used as the barrier. The skin from the inner part of the ear of a rabbit (6 months old) was obtained from a local slaughterhouse, frozen and used within 30 days.  

 PVA was hydrated in the appropriate volume of water to a final concentration of 20% (w/w). PVP was dissolved (21% w/w) in 15 mL of a mixture of polyethylene glycol (PEG) 400:water (15:85 w/w) and the appropriate amount of bupropion or nicotine was added. This solution was mixed with 80 g of PVA solution under continuous stirring. The resulting mixture was spread on siliconized paper (Rexam Release; Apeldoom, NL, USA), using a film-casting knife (BYK Gardner; Silversprings, MD, USA) with a gap of 400 μm, and then left to dry at room temperature overnight. Two different combined films (A and B) were prepared using a fixed concentration of nicotine and two different concentrations of bupropion.  

Once dried, three circles (26 mm in diameter) were cut from each film. Each circle was weighted and dissolved in 100 mL of a solution of 0.1 N HCl. The solutions obtained were analysed by HPLC to determine the amount of nicotine and bupropion contained in the patch. The results were expressed as %(w/w) and as mg/cm

 HPLC analyses were performed with validated methods, using a Perkin Elmer instrument (Norwalk, CT, USA) and a Novapak C18 column (Waters; Milford, MA, USA). The UV detector was set at 254 nm.  

The mobile phase for nicotine was a mixture of 1 M CH

COONa, 0.25 M sodium dodecyl sulphate, water and CH

CN (1:1:68:30 v:v) pumped at 1.5 mL/min.  

The mobile phase for bupropion was methanol (40% v/v) and 0.1 M sodium acetate containing 0.01 M tetrabutylammonium bromide to pH 5.0 (60% v/v), pumped at 1.0 mL/min. 

 Permeation experiments were conducted in Franz-type diffusion cells (Disa; Milan, Italy) with an exposed surface area of 0.6 cm

 . Rabbit ear skin was used as membrane and the receptor phase was 0.9% NaCl solution. The donor solutions tested were: 1 mg/mL solution of bupropion HCl in pH 7.4 HEPES buffer; 1 mg/mL solution of nicotine in pH 7.4 HEPES buffer; nicotine and bupropion hydrochloride (both 1 mg/mL) solution in pH 7.4 HEPES buffer; nicotine and bupropion containing transdermal films Table 1); and Nicorette commercial patch (Pharmacia & Upjohn; Milan, Italy) (Table 1). 

 The profiles obtained from the solutions of bupropion and nicotine were analysed with Equation 1, a solution of Fick's second law of diffusion in a homogeneous membrane:

 where Q(t) is the amount of drug permeated as function of time, K is the partition coefficient, H is the diffusional path-length, C

 is the donor concentration and D is the diffusion coefficient. 

The fitting was performed with n equal to 10, using Kaleidagraph software (Synergy Software; Reading, PA, USA) on a Power G3 MacIntosh computer (Apple Computer; Cupertino, CA, USA). The permeation parameters (KH and D/H

 ) calculated for the two molecules are presented in Table 2. KH gives an indication of the partitioning of the molecule between the 

  represents the diffusive parameter. The permeability coefficient (P) was calculated as KH×D/H

The results were expressed as the mean±sem. Statistical differences were determined by the Student's t-test, which assess whether the means of two groups are statistically different from each other. 

 At first, the transdermal penetration of bupropion and nicotine was studied individually, starting from 1 mg/mL aqueous solutions at pH 7.4. Both nicotine and bupropion are basic drugs with a pKa of 7.9 (Table 2). At the pH of the experiment, 76% of each drug was in its protonated form (the value estimated from the Henderson-Hasselbach equation). The permeation profiles obtained are presented in Figure 1 (bupropion) and Figure 2 (nicotine). Bupropion showed good permeation characteristics and very short lag time, but after approximately 6 h the permeation rate seemed to decrease, probably as a result of the depletion of the donor compartment. After 24 h, approximately 60% of the amount applied had permeated the skin. The profile of nicotine permeation remained linear up to 24 h (Figure 2). In this case, the percentage permeated after 24 h was approximately 10%. Fluxes were calculated as the slope of the linear portion of the curves and are shown in Table 2. 

As the aim of this work was to formulate nicotine and bupropion in a single transdermal film, the next step was to evaluate their reciprocal influence on transdermal permeation. This was done by measuring the transdermal flux of each drug from a solution containing both. The profiles obtained (Figures 1 and 2) showed a slight increase for both permeants. This increase is not statistically significant and suggests the absence of any reciprocal influence on their transdermal penetration, as confirmed by the flux data reported in Table 2. 

The permeation profiles of both molecules were then analysed to calculate the relevant parameters, thanks to an appropriate solution of Fick's second law of diffusion in a homogeneous membrane (Equation 1, applied 0–24 h for nicotine and 0–6 h for bupropion). The permeation parameters (KH and D/H

The permeability coefficient of bupropion was higher than that of nicotine, probably because of the higher lipophilicity of bupropion, as demonstrated by its logD value (Table 2). The permeability coefficient of nicotine at pH 7.0 reported in the literature across human skin (8.4×10

  cm/h) is in agreement with our data on rabbit ear skin,

  which confirms that rabbit ear skin can be a reasonable model for human skin.

  To our knowledge, no data on bupropion transdermal permeation starting from a solution are present in the literature. 

As a reference to discuss bupropion data obtained in the present work and the feasibility of bupropion transdermal administration, the minimum amount of drug necessary to maintain therapeutic plasma levels was calculated. From the total clearance (133 L/h) and the concentration at steady state (136 μg/L), the input rate was calculated as 18 mg/h.

 Considering an application area of 20 cm

 , and the permeability coefficient obtained in our study (0.0284 cm/h), a donor concentration of approximately 30 mg/mL would be necessary to achieve the therapeutic plasma concentration. It should be kept in mind that this calculation is only a rough estimate of the feasibility of bupropion transdermal delivery, considering that our data was obtained 

 Initially, the permeation of nicotine was studied from the reference market product Nicorette. The commercial product tested is a drug in adhesive patch, containing nicotine at the dose of 8.3 mg/10 cm

 release is 5 mg/16 h, that is, 500 μg/cm

  over 16 h. The permeation profile obtained (Figure 3(a)) is not linear with time, but, as previously reported,

  decreases with time. This suggests a diffusional control of the polymeric matrix on nicotine release. The release rate obtained by plotting the amount permeated against the square root of time was 129 μg cm

 , which agrees with the published data across human epidermis (119 μg cm

The amount permeated at 24 h was ≈600 μg/cm

  and represents 70% of the amount loaded, which is a surprisingly high percentage for a transdermal patch. Although effective, the transdermal systems containing nicotine show local adverse reactions that occur in 30% of patients. Reactions include irritation induced by occlusion; irritation or allergy to the delivery system (matrix and adhesive layer); and erythema or allergy caused by nicotine.

To try to reduce these problems, we developed an innovative transdermal drug delivery system, a water-based and vapour-permeable bioadhesive film intended for transdermal delivery (Figure 4).

  The transdermal film called Patch-non-Patch is not adhesive in the dry state, but is adhesive when applied to wet skin and is also water permeable, which reduces the risk of skin irritation. The typical composition includes a film-forming agent (PVA), an adhesive (PVP) and a plasticizer (PEG 400). The water solution of all the components is spread on silicone paper and oven-dried to produce the final film. Film application requires preventive skin wetting to partly dissolve the adhesive and restore skin adhesiveness. 

The skin adhesion properties of the film have been previously characterized.

  The value of adhesion force was in the range 0.6–1.8 N, while the work of adhesion was 9–30 mJ. The values of adhesion force were not significantly modified after 24 h of application on the skin 

 and the film was still firmly attached to the skin. 

Two different combined films were prepared (A and B), using a fixed nicotine concentration (similar to the commercial product), but with two different bupropion concentrations (Table 1). The permeation profiles obtained from the films are reported in Figure 3(a) and Figure 5. The nicotine permeation profile (Figure 3(a)) showed matrix-type kinetics and was comparable to the one obtained with the commercial product. However, the film presented a more pronounced initial 'burst' effect. The average nicotine flux in each sampling interval was calculated for the film and for the commercial patch, and the result is reported in Figure 3(b). 

The two formulations produced different fluxes during the first two hours of the experiment, while thereafter the profiles were almost superimposable. In particular, the flux from the film was much higher during the first two hours, indicating the possibility to achieve a faster appearance of nicotine in plasma when administered 

. However, it should be underlined that a higher flux has a higher irritation potential because nicotine is skin irritant. 

Nevertheless, the film prepared is water-based and non-occlusive, and these properties should lower the irritation potential (although only the 

 skin irritation test can give the definite answer). As a result of the non-occlusivity of the film and to the volatility of nicotine, the partial vaporization of nicotine at the air interface during the wear period can be envisaged. This should be taken into account for future development of the product as the use of a non-occlusive backing can lead to a reduction of nicotine concentration because of volatility, thus reducing the performance of the film. 

The permeation of bupropion from the two different films, A and B (Table 2), is reported in Figure 5. Using the low concentration patch, bupropion flux was quite low and the percentage permeated reached only 11% of the amount loaded. The increase in drug loading determined a proportional increase in the amount permeated, but the percentage permeated was not substantially different (15%). 

In any case, the profiles are characterized by the absence of lag-time and by the presence of an initial 'burst' effect that could be useful for faster onset of the therapeutic effect. The low percentage of bupropion permeated from the patch is in disagreement with the good permeability characteristics shown by bupropion in solution. One possible explanation is that bupropion in the patch is present entirely in its ionized form (the pH of the patch is approximately 3.5). Owing to the low permeation rate, the requested input rate (6 mg/h) could not be achieved with the prepared films. However, because of the proportionality between the amount loaded and the cumulative amount permeated (observed also by Gondaliya 

 ) a therapeutically useful permeation rate can potentially be obtained by increasing drug loading. 

Both bupropion and nicotine showed good permeation characteristics when applied to the skin in solution at pH 7.4, but the combination of the two substances did not significantly modify their transdermal flux.  

The inclusion of the two substances in an innovative transdermal film resulted in particularly interesting findings because a high transdermal flux in the early times of patch application was obtained, suggesting a quick onset of the relevant plasma levels 

. In the case of nicotine, the flux obtained at 2 h was significantly higher than that obtained from the commercial patch. The fluxes obtained for bupropion are not able to generate therapeutic plasma levels, but the results suggest that the required flux could be achieved by increasing drug loading. 

Articles

Transdermal film containing nicotine and bupropion for combined smoking cessation therapy

Cleaning may be defined as the physical or chemical removal of undesirable substances from apparatus surfaces. All substances remaining in the finished product or within processing equipment are contaminants.1 The current good manufacturing practice (cGMP) regulations emphasize that cleaning is a critical issue to ensure product quality.2 In response to compliance and other manufacturing and quality related issues, clean-in-place (CIP) systems and technologies have been applied to pharmaceutical processing equipment to facilitate validated cleaning.	

A continuous movement capsule filling machine (Imatic 150; IMA SpA, Bologna, Italy) was used for this investigation, which utilizes a system that allows all filling operations to occur within a single turret with all capsule movement being guided mechanically and smoothly. Features of the machine include maximum dosing precision and reliability, even at high speeds. In this study, the production area was totally isolated from the mechanical area, and the machine was fitted with a completely automatic CIP system to reduce manual intervention. The cleaning programme resulted in a clean, dry, ready-to-use machine. 	

The filling machine and automated CIP system 

The Imatic 150 consists of a main frame, standing on antivibration feet, with rounded edges that supports the rotating turret and the machine drives. The main frame has an AISI 304 stainless steel cover, on which the machine enclosure is mounted. The operating area within is completely isolated from the drive and any mechanical movement located inside the main frame.				

It is a continuous motion machine with the main turret rotating in a clockwise direction at a maximum speed of 70 rpm. The bushes that carry the caps and the body of the capsules are located at each operating station of the turret and are specified according to the size of capsule to be filled.	

As the turret rotates, the following operations are performed on the capsules:	

feeding, orienting and opening the capsules (with the capsules loaded into their hopper either manually or by using an SC1000 elevator)

product dosing by means of a tubular volumetric dosing system, drawing the powder from a rotating (or vacuum) type bowl

identification and rejection of defective capsules

ejection of the closed capsules and sampling

A statistical checkweighing operation (if required) can be performed by an optional unit, which produces batch production records. Machine control and supervision are performed using a single, industrial PC. The operator interface is positioned next to the machine and is divided into two parts: 	

the upper part houses a display touchscreen, a membrane keyboard, start/stop switches and an emergency switch

located in the lower part are the optional groups for production control (such as a statistic weight checking unit and a quality control sampling unit).

The processing area was completely separated from the mechanical area by the use of Viton seals (Forsheda AB, Forsheda, Sweden) and polyurethane bellows (TSE Industries, Clearwater, Florida, USA). The seals were used to isolate the rotating turret from the bowl section and from the machine base, whereas bellows were applied to the following moving parts: pushers for opening/closing capsules, the capsule body transport shaft, the orientation device shaft, the dosing system and piston shafts.	

This separation prevented the passage of powder into the mechanical area, thus limiting the surfaces to be washed, and allows the machine to be fitted with a completely automatic CIP system, preventing the passage of water into the mechanical area. At the end of the production process, the operator prepares the machine to be washed, as follows:	

open the doors and, using an aspirator, empty the machine of empty capsules and residual powder

dismantle the empty capsule feeding channels, powder bowl and empty capsule hopper

remove the powder feeding tube and replace it with a cover with a spray ball (the capsule feeding tube being automatically closed)

plug all holes (including capsule exits) and complete the assembly of the cleaning devices (some of which are always fitted on the machine), then close the doors and start the cleaning cycle.

Washing devices consist of fixed-blade spraying nozzles and a rotating spray ball: they are positioned within the turret so that both the processing area and the turret (rotating at a low speed during the washing cycle) are cleaned. The processing area is provided with a draining hole to collect the sprayed water.	

The evaluation of the CIP procedure was done according to PDA Technical Report No 29. Points to Consider for Cleaning Validation.3 In particular, the validation was done using the following trials:	

Cleaning trial: rinse and swab sampling after the production of a capsule batch containing a tracer, to quantify the presence of the tracer in the internal part of the machine and in the discarded water.	

Contamination trial: sampling of a placebo capsule batch produced after the cleaning procedure to check for possible tracer contamination.	

Recovery trial: rinse sample to check whether the amount of added tracer would be totally recovered during the cleaning procedure.	

To assess the effectiveness of a cleaning process, cleaning validation procedures are typically performed using a tracer. The process is considered effective if a non-significant amount of tracer is found in the cleaned machine or in the subsequently prepared product batches. 	

To address this issue, both acidic and basic detergents are usually employed, which solubilize the contaminants so that they are eliminated with the rinsing water. However, the capsule filling machine in this study is made of aluminium. Thus, it can only be cleaned with acidic detergents because caustic chemicals corrode the metal. Therefore, the selection of an acidic tracer (fluorescein) that cannot be solubilized by the detergent may represent the 'worst' and most critical condition for the process validation. It was considered that if the cleaning procedure was effective in this situation, there was a good possibility that it will be successful in many other, less testing conditions.	

Cleaning trial. The filling mixture (60 kg), containing 68% w/w spray-dried lactose (Eingemann & Veronelli, Milan, Italy), 30% w/w wheat starch (Eingemann & Veronelli), 2% magnesium stearate (pharmacopoeia grade) and 0.5% w/w sodium fluorescein (Fratelli Fiorio, Milan, Italy) as the water-soluble tracer, was prepared with a cylindrical mixer by applying the "doubling-up" technique to ensure an even distribution. A batch of capsules, each containing 400 mg of filling mixture, was prepared by running the machine for 1 h at 150000 cps/h. Fluorescein was selected as the tracer for the aforementioned reasons and because it is detectable at a very low concentration. After production, the machine was prepared for washing using the previously mentioned operations. The cleaning programme was developed through the following stages:				

Pre-washing: hot tap water (56 Â°C) was pumped at a pressure of 5 bar and a flux of 100 L/min through all the nozzles. This stage aimed to remove most of the residual powder and was done in three stages (P1, P2 and P3). Sprayed water was collected at the base of the turret and drained through the draining hole. The total amount of water used during the pre-washing procedure was 417 L.	

Washing: hot tap water (56 Â°C) containing 0.7% v/v of an acidic detergent (Antos PH; DK Ausiliari Chimici, Bologna, Italy) was pumped at the same pressure and flux used in the pre-washing stage. The detergent contained phosphoric functional groups; thus, the efficacy of the rinsing procedure could be assessed by analysing the phosphate content of the rinse water discarded from the machine. This stage, meant to completely remove soluble and insoluble residuals of the production process, was performed in three steps (W1, W2 and W3) and used 415 L of water.	

Rinsing: a total volume of 571 L of water was pumped through all the nozzles to completely remove the detergent. This procedure was done in two phases, each comprising two steps: 	

Phase 1: 288 L of hot tap water (56 Â°C) were pumped at 5 bar and 100 L/min through steps R1 and R2

Phase 2: 283 L of demineralized water (20 Â°C) were pumped at 7 bar and 100 L/min through steps R3 and R4 to remove all residual salts and soluble substances.

Drying: A coarse dripping was performed by rotating the turret at 70 rpm for 30 min. The drying phase was then performed by insufflating hot air (70 Â°C) for 1 h through both the water-spraying nozzles and the draining hole.	

Sample collection. For evaluation purposes, three samples (8 mL each) of the solution/suspension of each stage of the pre-washing, washing and rinsing stages, respectively, were collected. In total, 30 sampling collections were performed. After the drying procedure, the machine-washed surfaces were visually inspected both under ambient and UV (253 nm) lighting.	

Quantification of fluorescein on the surfaces was made by swabbing them with a hydrophilic cotton pad (surface of 24 cm2) wetted with 1 mL of 0.5% w/v solution of NaCO3 in water. Table I reports the parts of the machine swabbed and the relevant surface. Afterwards, the pads were extracted with a known volume of 0.1 M phosphate buffer solution pH 8,4 filtered and analysed for fluorescein content. The extraction procedure was previously validated (mean recovery599.92%).	

Sample preparation and analysis. All samples were analysed for pH and, after dilution with an appropriate volume of pH 8 buffer, for fluorescein content. The obtained solution was filtered through a 0.45 Âµm membrane (Millipore Corp., Billerica, Massachusetts, USA). Fluorescein determination was performed spectrophotometrically at 493 nm, and spectrofluorimetrically at 515 nm. The lower limits of quantification were 0.02 and 0.001 ppm, respectively. Rinse water samples were also analysed for total organic content (TOC)5 and phosphate content.6	

Upon completion of the cleaning trial, a batch of empty capsules was produced to check the possible contamination of the capsule shell. Then, to verify this, a second batch of placebo capsules was produced using a mixture that had the same composition as the filling mixture, but without tracer.	

Table I:  Area and fluorescein per unit area of the different parts of the machine swabbed after the CIP procedure.				

Ten empty capsules were dissolved in 20 mL of pH 8 buffer (50 Â°C). After filtration, the samples were analysed for fluorescein content. The content of 10 placebo capsules was dispersed in 20 mL of the pH 8 buffer. The obtained suspension was also filtered and analysed for fluorescein content.	

The aim of the recovery trial was to verify that the tracer (purposely deposited into the machine) could be completely removed during the cleaning procedure and could, therefore, be totally recovered in the discarded water. Thus, after dismantling, 300 g of a mixture containing 6% w/w spray-dried lactose (Eingemann & Veronelli), 39% w/w microcrystalline cellulose (Avicel PH 102, Eingemann & Veronelli), 1% w/w magnesium stearate (pharmacopoeia grade) and 0.5% w/w fluorescein, were manually dispersed into the working parts of the machine to simulate normal production dusting. A complete cleaning process was performed and the water of each step was collected in a tank; NaOH was added to each tank to increase the pH to 9 to solubilize fluorescein. After thorough mixing, samples were collected from each tank, filtered and analysed for their fluorescein content. 				

The objective of the cleaning validation was to verify the effectiveness of the cleaning procedure to remove product residues or cleaning agents. The criteria adopted for cleaning evaluation were those presented in the "Cleaning Validation Guidelines"7 which, at present, can be considered the most restrictive guideline among those edited by the regulatory authority of the different countries. The criteria were		

Figure 1: Fluorescein concentration in discarded water of the various steps of pre-washing, washing and rinsing stages. The horizontal line indicates the limit of quantification of the analytical method used and Figure 2:  pH values measured in the discarded water of the various steps of pre-washing, washing and rinsing stages.				

not more than 0.1% of dose of any product should appear in the maximum daily dose of following product

not more than 10 ppm of any product should appear in another product

The aim of this study was to validate the automated clean-in-place (CIP) system installed on a capsule filling machine to determine its ability to adequately eliminate contaminants. The results obtained from the proposed cleaning validation trial showed that all the soluble tracer was removed after the washing procedure. At the end of the CIP procedure, the discharged water had the same pH, phosphate content and total organic content as the supplied water. Lack of cross-contamination in the product was also demonstrated and a recovery trial highlighted the complete elimination of the tracer from the machine.