Roberto Trebbi

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

Articles

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.