Articles

Issue PDF

Implementing Toyotaâ€™s LEAN Systems in Bioprocessing

BioPharm International, October 2009 Issue (PDF)

Traditionally, the summer months tend to be quiet for the biotech sector. However, the general markets experienced three solid months of gains as investors started to believe that the worst of the economic woes were over. This upbeat mood spilled over to biotech with the Burrill Biotech Select Index posting a solid gain of almost 11% from the June to August period, staying pace with the Dow, up 11.7% and NASDAQ, up 13.23% during the same period. 

Adding to the favorable capital markets, biotech impressed investors with positive drug data, strong drug sales and earnings, and partnering and M&A deals. Shares of Human Genome Sciences, for example, jumped a whopping 690% after the company announced that its drug candidate Benlysta reduced lupus symptoms. Some of the share price increase was also driven on rumors that GlaxoSmithKline (GSK) could acquire Human Genome Sciences because GSK has collaborated on Benlysta development. 

Investors also rewarded Savient Pharmaceuticals whose shares soared 119% after the FDA said its drug Krystexxa appeared to successfully relieve swollen joints and pain flare ups associated with gout. Other major biotech movers in the June to August period included:  

Targacept, which closed up a whopping 424% mainly on news that its drug candidate, TC-5214, significantly improved symptoms of major depressive disorder for patients who did not respond to an older drug, Forest Laboratories Inc.'s Celexa. 

Orexigen Therapeutics' shares jumped 122% after revealing that Contrave, a longer lasting form of two generic diet drug compounds, helped people lose weight in all three of its major clinical trials. That should clear the way for the company to file for approval to market the drug with the FDA.

Medarex's shares soared 90% in July on news that Bristol-Myers Squibb (BMS) was buying the company for $2.4 billion, or $16 per share in cash, representing a 90% premium. With the acquisition, BMS gains full rights to ipilimumab, an experimental drug for metastatic melanoma in late-stage testing. BMS also gains rights to seven antibodies in clinical trials under Medarex's sole sponsorship and three other antibodies being co-developed with other partners, and rights to preclinical assets in various stages of development by Medarex.

In addition, we did see an IPO get out the door in August. Cumberland Pharmaceuticals raised $85 million by offering 5,000,000 shares at $17, below the expected range of $19 to $21. The specialty pharmaceuticals company has products on the market, including a new injectable treatment for pain and fever.  

The Cumberland IPO will not open the biotech IPO window just yet, but companies on the runway with products already on the market, or close to it, might be tempted to test the waters. Early-stage biotechs, however, are unlikely to find favor in a climate that still has no tolerance for risk. 

An international IPO was also completed during August. Israeli biotech company, D-Pharm, listed on the Tel Aviv Stock Exchange, raised $7.6 million in an oversubscribed offering. The offering price of approximately $4.50 per share was higher than D-Pharm's original minimum goal, but still considered a bargain considering the company's pre-IPO valuation of about $32 million. D-Pharm also raised $15.4 million in a rights offering at the same time and at the same terms. The company will use most of the proceeds to fund a late-stage clinical trial of its lead drug DP-b99 in moderate to severe ischemic stroke for which the only currently approved drug is tissue-plasminogen activator (t-PA), marketed by Genentech as Activase.  

Our Burrill Report recorded a significant 25% increase in total financings and partnering in Q2 '09 compared to Q1 '09 (Table 1). Although financings were down quarter over quarter, the amount US biotech firms raised in partnering deals was up 68%. 

Table 1. Quarter comparisons of US biotech financings (USD) 

This increase shows that in a tight fiscal environment, companies are devoting their energies on finding pharmaceutical and biotech partners to help build value of their lead product programs. It has become increasingly difficult to attract venture capital (VC) and this is reflected by the fact that the amount raised by biotechs in the US in the second quarter was 43% down from the total raised in the first quarter. 

Table 2. Year-to-date (YTD), the Burrill Biotech Select Index was up 3% 

At the end of August, the group of public biotech companies had an aggregate market cap of $340 billion (Table 2). Statistics also show that:  

there are 49 companies with market caps greater than $1 billion

there are 36 companies that have market caps between $500 and $999 million 

a total of 136 public biotech companies (41%) have a market cap below $100 million 

approximately $530 million in VC was raised by US biotech companies in August. Notable deals included: Pacific Biosciences ($65 million), Complete Genomics ($45 million), and Adamas Pharmaceuticals ($40 million).

year-to-date, the Burrill Biotech Select Index is up 3%.

Biotech is not yet fully back on track because many companies are still struggling to find the necessary funding to maintain their operations. Almost half of US public biotechs have market caps below $100 million and we are seeing companies still shutting shop. It is important to remind ourselves that the biotech industry is undergoing a major transition, a process that will likely continue for many months. 

This is because we do not know how President Barack Obama's proposal for healthcare reform will fully impact the biotechnology industry and the status of biosimiliar legislation is also still unresolved. There are fears that these issues will drive the prices of innovative drugs lower and eat away at biotech company profits. 

Going into the final quarter of the year, which is traditionally one of biotech's best, we will likely see the industry continue to build on its momentum gained during the past several months. 

 is chief executive officer at Burrill & Company, San Francisco, CA, 415.591.5400, 

Burrill on Biotech

Biotech impressed investors with positive drug data, strong drug sales and earnings, and partnering deals.

Biotech Remains Upbeat During the Summer

The year 2008 was, like 2007, a somewhat disappointing year in terms of the number of new biopharmaceuticals approved. The year witnessed the approval of eight recombinant proteins or engineered antibody-based products in the US, and two such biopharmaceuticals in the European Union (EU, Table 1). These products included hormones, interferons, blood factors, and fusion proteins—two each—as well as one antibody-based product and a colony-stimulating factor. 

Although 10 approvals does not appear an unduly low number, only four of those products contained genuinely new active ingredients (Arcalyst, Cimzia, Nplate, and Recothrom; Table 1). Arcalyst and Nplate are fusion proteins, Cimzia is an antibody Fab', and Recothrom is a recombinant thrombin (blood coagulation factor IIa).

The other six (Accretropin, Extavia, Filgrastim ratiopharm, NovoLog mix, PEGintron/ribetol combo, and Xyntha) are effectively reformulated, biosimilar or "me too" versions of previously approved products (see sidebar: 

Product indications too are somewhat diverse and none aim to treat cancer. Also notable amongst the 2008 crop is the paucity of antibody-based products. No full-length antibody came to the market last year, although one antigen-binding antibody fragment (Cimzia) did gain approval. 

In terms of expression systems used, the 2008 approvals also appeared to buck the general trend toward mammalian-based systems. Six of the 10 approved products are expressed in 

, and three in Chinese hamster ovary (CHO) cell lines. However, the 

 bias is less obvious in the context of genuinely novel active pharmaceutical ingredients approved, of which two are expressed in 

In most previous years too, some overlap in EU and US approvals was evident.

  Not so in 2008. Although approval numbers in the US were modest, they were even more disappointing in EU, with only two products coming on the market. Moreover, neither of the EU approvals (Extivia and Filgrastim Ratiopharm) were genuinely novel (see sidebar). 

In financial terms too, the majority of products approved, are expected to generate modest as opposed to blockbuster-level revenues. Nplate, Xyntha, and Recthrom sales are each projected to reach the $300–500 million mark.

  Arcalyst, aimed at treating CAPS (Table 1) has designated orphan status. With an estimated total US population base of only 300–500 patients, the market potential has been estimated at $10–35 million, despite the drug's premium price (approximately €250,000 per patient annually).

Table 1. Biopharmaceuticals (defined as recombinant proteins, monoclonal antibody, and nucleic acid-based products) approved in the US or EU in 2008 

The only product with a likely potential of reaching blockbuster status is that of Cimzia.

  Although initially approved only for the treatment of Chron's disease, it was subsequently approved (May 2009) for the treatment of rheumatoid arthritis, greatly increasing its revenue scope.  

Overall, therefore, 2008 was a modest year in terms of biopharmaceutical approvals. Analysts, however, are hopeful that this will represent a low point, with approvals likely to steadily increase once again in the coming years.

Individual monographs for products are presented below. Monograph for Extavia, Novologmix 50:50, and PEGintron ribetol are not provided, given their particularly close similarity to well-established, previously approved products.  

 is a recombinant form of human growth hormone (hGH), expressed in 

. Native hGH is a 191 amino acid, 22kDa polypeptide hormone synthesized in the anterior pituitary and Accretropin is identical to the native molecule in amino acid sequence. Accretropin is a biosimilar-type product. It was approved in January 2008 under section 505 (b) of the food, drug, and cosmetic act—a similar route under which Sandoz's hGH product, Omnitrope, gained approval in 2006. 

Accretropin is indicated for the treatment of pediatric patients displaying growth failure because of inadequate secretion of endogenous hGH and for short stature associated with Turner syndrome (a chromosomal disorder) in pediatric patients whose epiphyses (the rounded end of long bones) are not closed. Dosage regimes should be individualized for each patient, but normally entail daily subcutaneous administration of 0.18–0.36 mg/kg body weight. Safety and efficacy were assessed in two trials containing 44 patients with growth hormone deficiency and 37 patients with Turner syndrome, respectively. Clinical results confirmed accelerated growth (height velocity) and the most common adverse reactions related to injection site events including erythema, edema, and brusing.  

Product manufacture is initiated by fermentation of the engineered 

. As initially synthesized, the recombinant product displays an additional N-terminal methionine residue, a consequence of the expression construct used. Subsequent to product recovery the additional methionine is cleaved, yielding a product identical in sequence to the native hormone. Additional downstream processing entails multistep chromatographic purification and sterile filtration. Final formulation entails the addition of NaCl, phenol, pluronic F-68 surfactant, and phosphate buffer constituents as excipients, and the product is presented in liquid solution (1 mL of a 5 mg/mL solution in glass vials). Accretropin is manufactured by Cangene corporation (Winnipeg, Canada) and is marketed in the US by Apotex, Cangene's majority shareholder.  

 is an engineered dimeric fusion protein with each of its two constituent polypeptides consisting of the ligand-binding domains of the human interleukin-1 receptor (IL-R1) and interleukin 1 receptor accessory protein (IL-1RacP), linked to the constant (Fc) portion of human IgG1. The 251 kDa glycosylated product is expressed in an engineered CHO cell line. It was approved in February 2008 by the FDA for the treatment of cryopyrin-associated periodic syndromes (CAPS), including familial cold autoinflammatory syndrome (FCAS) and Muckle-Wells syndrome (MWS), in persons 12 years and older.

Autoinflammatory syndromes are usually underlined by aberrant regulation of cytokine signaling. CAPS (including FCAS and MWS) encompass one such group of rare autoinflammatory conditions, representing a spectrum of one disease manifesting with varying degrees of severity. The conditions are linked to mutations in a gene coding for a protein known as cryopyrin, which plays a role in generating free biologically active interleukin-1  (IL-1β). As a result, overproduction of IL-1β occurs, driving inflammation. General features of these conditions include recurrent fevers, rash, and joint pain.  

The discovery of the molecular basis of this disease group suggested the blocking of IL-1β activity as a potential treatment. Some CAPS patients were thus treated with Kineret (anakinra, an IL-1 receptor antagonist approved for the treatment of rheumatoid arthritis since 2001), although that product's short half-life necessitated daily injections. Arcalyst also brings about its therapeutic effect by acting as a decoy receptor, binding excess free IL-1β, and hence, preventing it from interacting with cell-surface receptors. The product's antibody Fc portion confers on it an extended half-life, supporting once weekly injections. 

Arcalyst's safety and efficacy profiles were established by a randomized, double-blind, placebo-controlled study involving 47 CAPS patients. Participants rated primary signs and symptoms of CAPS (joint pain, fever or chills, rash, eye pain or redness, and fatigue) on a scale of severity from 1–10, with Arcalyst-treated patients reporting a mean reduction in symptom severity of –2.4. Moreover, mean symptom scores subsequently increase when patients were withdrawn to placebo. The most common adverse reactions reported were injection-site reaction and an increased susceptibility to infections.  

Arcalyst is manufactured and distributed by Regeneron Pharmaceuticals (New York, NY). Manufacture entails initial cell culture of the CHO producer cells, followed by chromatographic purification. The product is marketed in a lyophilized form and contains histidine, arginine glycine, sucrose, and polyethylene glycol as excipients.  

 is a recombinant humanized anti-TNF-α antigen-binding (Fab') fragment covalently linked to a 40kDa polyethylene glycol. The 91 kDa antibody fragment is expressed in 

 and is composed of a 214 amino acid light-chain and a 229 amino acid heavy-chain fragment. The FDA approved the product in April 2008 for the treatment of Crohn's disease in adult patients who have not responded adequately to conventional therapy. 

Crohn's disease is a chronic inflammatory condition of the gastrointestinal tract. Characteristic symptoms can include abdominal pain, diarrhea, vomiting, or weight loss. The condition is characterized by overactive T cell activity, including the overproduction of proinflammatory cytokines—most notably tumor necrosis factor-α (TNF-α). Cimzia brings about its therapeutic effect by binding (mopping up) TNF-α, thereby reducing the pro-inflammatory effect. Unlike additional full-length antibody-based therapies for Chron's disease (Remicade and Humira), Cimzia is devoid of the antibody Fc component, and therefore, cannot trigger Fc-mediated functions, such as complement fixation or antibody-dependant cell-mediated cytotoxicity. The PEG moiety extends the product's plasma half-life (antibody fragments are characterized by substantially shorter half-lives than intact antibodies), allowing an initial dosage regimen of once fortnightly subcutaneous injection, with subsequent once monthly maintenance dosages.

Product safety and efficacy was primarily established in two double-blind, randomized placebo-controlled studies with a combined population of over 1,000 Crohn's sufferers. Overall, the studies confirmed a statistically significant improvement in product-treated patients as compared to the placebo group. The most common—and potentially life threatening—adverse reactions were increased risk of infection, with tuberculosis, invasive fungal, and other opportunistic infections, all being observed in Cimiza-treated patients.  

Manufacture entails chromatographic purification subsequent to product recovery from 

-based fermentation and the final product is formulated in a lyophilized form, containing sucrose, lactic acid, and polysorbate as excipients.  

 is a recombinant form of the 174 amino acid native human granulocyte colony stimulating factor (G-CSF) expressed in 

. Compared to native G-CSF the product is nonglycosylated and contains an extra N-terminal methione reside, both modifications reflecting the recombinant expression system used. Approved for general medical use throughout the EU in September 2008, Filgrastim ratiopharm is a biosimilar product and the reference product against which it was compared was Amgen's Neupogen (also produced in 

, also nonglycosylated and of an identical amino acid sequence).  

Filgrastim ratiopharm is indicated for reducing duration or incidence of neutropenia associated with certain conditions, including cancer therapeutic regimes, advanced HIV infection, and congenital neutropenia. Neutropenia is a condition triggered by a significant decrease in blood neutrophil (white blood cells capable of ingesting and destroying bacteria) levels and is characterized by the occurrence of frequent and usually serious infections. G-CSF functions as a growth and differentiation factor for neutrophils and their precursor cells, and activates mature neutrophils. Clinical comparability to Neupogen was investigated in one main trial involving 348 breast cancer patients suffering from severe chemotherapy-associated neutropenia and with follow-up safety assessments in patients with non-Hodgkin's lymphoma and lung cancer. Safety and efficacy were comparable to the reference medicine, with muscoskeletal pain representing the most common side effect recorded.  

The active substance is manufactured by SICOR Biotech (Vilnius, Lithuania) with Ratiopharm (Ulm, Germany) holding the marketing licence. Manufacture entails initial fermentation of the producing 

, induction of product synethesis, and subsequent recovery of product-containing inclusion bodies from the harvested cells. The inclusion bodies are dissolved using a chaotrophic agent, with protein refolding then undertaken in a reducing–oxidizing system. The product is then purified to homogeniety by column chromatography. The final product contains sorbitol, polysorbate, and sodium acetate buffer components as excipients and it is presented in solution in prefilled syringes.   

 is a dimeric fusion product. Each of its two constituent polypeptides consists of two thrombopoietin receptor-binding domains linked to a human IgG-1 Fc fragment. It is expressed in 

. Approved by the FDA in August 2008, Nplate is indicated for the treatment of thrombocytopenia (low platelet count and hence impaired blood clotting ability) in patients with chronic idiopathic thrombocytopenic purpura (ITP) who display insufficient response to the alternative potential treatments of corticosteroids, immunoglobulins, or splenectomy. ITP is an autoimmune condition characterized by the destruction of platelets (thrombocytes) and megakaryocytes (the platelet-producing cells). 

Thrombopoietin (TPO) is the primary physiological (stimulatory) regulator of platelet production. Nplate, by its thrombopoietin receptor-binding domains, binds to and activates TPO cell surface receptors in a manner analogous to endogenous TPO. The resultant increase in platelet production counterbalances platelet destruction by the immune system. Nplate's IgG Fc component serves to increase the product's serum half-life, facilitating its once weekly subcutaneous administration. 

Safety and efficacy were largely assessed by two double-blind, placebo-controlled clinical studies and one open-label extension study involving 124 patients with chronic ITP. A clear overall increase in platelet production was observed in response to Nplate administration. The most common adverse effects noted included joint, stomach and shoulder pain, dizziness, and insomnia. However, several more serious effects were also reported, including a risk of bone marrow fibrosis and worsened thrombocytopenia after cessation of therapy.

Nplate is manufactured and distributed by Amgen (Thousand Oaks, CA). Subsequent to post-fermentation recovery and chromatographic purification, mannitol, sucrose, histidine, and polysorbate are added as excipients, and the final product is lyophilized.

Recothrom (thrombin, topical, recombinant)

 is a recombinant version of human thrombin produced in an engineered CHO cell line. It was approved by the FDA for controlling minor bleeding during surgery. The product is applied topically, directly on the surface of the bleeding or oozing tissue only. Thrombin (activated blood coagulation factor II, i.e., factor IIa) is a serine protease that functions at the terminal stage of the normal blood-clotting process. Its inactive circulating xymogen (prothrombin, i.e., factor II) is a 582 amino acid, 72.5 kDa glycoprotein containing six γ-carboxyglutamate residues toward its N-terminal end. It is proteolytically activated during the blood coagulation cascade by factor Xa and the active thrombin molecule consists of two polypeptides linked by a single disulfide linkage. The smaller polypeptide contains 49 amino acids while the larger contains 259 amino acids. [Physiologically thrombin then proteolytically activates fibrinogen (blood factor I), yielding fibrin, which forms the structural basis of a blood clot]. The recombinant thrombin displays an identical amino acid sequence to that of the native molecule.

The primary clinical evaluation undertaken was a multisite, randomized, double-blind study involving 411 surgical patients, who were treated with either recothrom or bovine thrombin. Efficacy was evaluated by the incidence of hemostasis 10 minutes post-product application, and no difference in efficacy between both treatment groups was recorded (95.4% versus 95.1%). The recombinant product, however, has a much reduced likelihood of triggering an immune response and the risk of accidental transmission of blood-bourne pathogens is also minimized. The most common side effect noted was complications at the incision site.

The recombinant molecule is synthesized and released from the producing CHO cell line in zymogen form. Downstream processing steps include its proteolytic activation, multiple chromatographic purification steps, and a solvent-detergent treatment step, as well as nanofiltration as viral clearance steps. No additives of human or animal origin are used in the manufacturing process. Histidine, mannitol, sucrose, sodium, and calcium chloride, as well as polyethylene glycol are added as excipients and the final product is lyophilized.

, PhD, is an associate professor in the Industrial Biochemistry Program at the University of Limerick, Ireland. He is also a member of 

 editorial advisory board, 353.61.202664, 

1. Walsh G. Biopharmaceuticals: approvals and approval trends in 2004. Biopharm Int. 2005;18,58-65.

2. Walsh G. Biopharmaceuticals: approval trends in 2005. BioPharm Int. 2006;9:58-68.

3. Walsh G. Biopharmaceuticals: approval trends in 2006. BioPharm Int. 2007;20:40-48.

4. Walsh G. Biopharmaceuticals: approval trends in 2007. BioPharm Int. 2008;21:52-65.

5. Raider RA. Low FDA approvals of new biopharmaceuticals continued in 2008. Available from: 

http://www.biopharma.com/approvals_2008.html

 6. Hoffman H, Yasothan U, Kirkpatrick P. Rilonacept. Nature reviews drug discovery. 2008;7:385-386.

Features

Ten biopharmaceuticals gained marketing authorization in the US or EU in 2008, although only four were new approvals.

Biopharmaceuticals: Approval Trends in 2008

The LEAN manufacturing and management model, developed from the Toyota Production System (TPS), has triggered major transformations in various manufacturing industries. Implementation in the field of biopharmaceuticals, however, has been limited to date. We believe that properly implementing LEAN in biopharmaceutical manufacturing can bring huge benefits, despite the complexity of biotechnology and the stringent regulatory requirements. In this article, we provide practical examples to highlight the areas where such benefits can be achieved. We also describe our qualification concept and the implementation of a new flat and crossfunctional process-oriented organization in manufacturing sites, in line with TPS principles, to ensure the development of a "LEAN" culture—a culture of continuous improvement.

LEAN can be defined as a manufacturing and management model that aims to reduce the time from customer order to delivery by eliminating sources of waste and by making the product flow through value-adding steps without interruption. It is derived from the Toyota Production System (TPS), which Toyota started developing in the 1950s. TPS has helped Toyota become the largest and one of the most profitable car manufacturers in the world. An excellent and comprehensive presentation of LEAN can be found in the literature.

  LEAN has since been implemented in many manufacturing industries, where it has triggered major transformations. However, LEAN has often brought only limited benefits. The first reason is that LEAN often has been implemented in a superficial manner, with a focus on "just-in-time" objectives only, without being understood as an entire system that must permeate an organization's culture.

  Furthermore, direct applications in the field of biopharmaceutical manufacturing have been quite limited to date because of concerns about the complexity of the associated technology and stringent regulatory requirements. A few nice examples, however, have been published recently.

We believe that implementing LEAN the proper way in biopharmaceutical manufacturing can bring huge benefits and help the industry to deal with increasing pressure on development and manufacturing costs, as well as with challenges in compliance and quality.  

In the first part of this article, we present two case studies in which LEAN principles have been applied successfully to large-scale biopharmaceutical manufacturing, using our radical methodology. In the second part, we describe our qualification strategy and the implementation of a new flat, process-oriented organization in manufacturing sites, to ensure optimal support of LEAN and the development of a culture of continuous improvement, as described by the TPS. We conclude with some additional opportunities for how LEAN can be applied more broadly in (bio)pharmaceutical companies, with great benefits. 

There are many different ways of implementing LEAN across an organization. Our methodology has been shaped over several years of application in the field of chemical, fill–finish, and now recently biomanufacturing operations. Our approach combines LEAN and the Six Sigma tools into a unique framework. The benefits of such a combination have been recognized by others.

1. The first step in our LEAN methodology is a two-day assessment workshop, starting with a simulation game to create a strong awareness of what can be achieved with LEAN. 

2. After defining the exact scope of the project, we assess the efficiency of the corresponding process—typically the manufacturing step including quality control (QC) and batch release activities—by drawing a high level "value stream map." We use only three measurements: throughput time, i.e., the overall cycle time for the production and release of a single batch; the throughput rate, i.e., the number of batches that can be produced per unit of time; and the failure rate, or "right first time" level. 

3. We then set the quantitative objectives of the project, using a radical approach. The team first defines the "blue sky vision" of the process, i.e., how it would look with none of the current constraints (e.g., regulatory, technological, organizational, economic, safety-related) (Figure 1). This blue sky vision corresponds to the ideal efficiency level. Then, only the constraints which cannot realistically be eliminated within the timeframe of the project are carefully added back. This approach leads to the "practical vision," which determines the LEAN objectives. Experience shows that the practical vision usually remains a very ambitious target with dramatic improvements compared to the current status (or "baseline"). In a few instances, it may even be equal to the blue sky vision. This approach has two main advantages over a traditional stepwise optimization process, where incremental improvements are made sequentially in different areas (e.g., technical, operational, organizational) or activities (e.g., manufacturing steps). First, with a clear vision of the end-stage, improvements typically are achieved faster and more dramatically. Second, the mental journey of going first to the ideal and theoretical situation, without claiming it as possible, turns out to be a very efficient way of circumventing the normal human resistance of team members during the assessment phase of a LEAN project. 

4. In the second day of the workshop, the LEAN implementation plan is developed, systematically including the following key milestones: 

a. selection and implementation of sensitive performance indicators 

b. determination of a "drumbeat" for batch production (typically on an hourly basis) 

c. overall equipment effectiveness (OEE) improvement at the identified bottleneck (usually the production bioreactor) 

d. synchronization of quality assurance (QA) and QC activities with production (thus minimizing waiting time for samples and accelerating batch record reviews) 

e. optimum sequencing of manufacturing activities 

f. throughput time reduction through the elimination of non-added value tasks  

g. implementation of a new process-oriented organization (see below). 

Several different technical tools are used to support LEAN, not only during the assessment workshop, but also during the execution phase of the project, such as a process walk, value stream mapping, spaghetti diagrams, rhythm wheels, failure mode and effects analysis (FMEA), and Six Sigma. (A description of these tools is beyond the scope of this article.) 

5. A multi-year implementation plan is then built from the above milestones, which typically are translated into specific sub-projects. Progress during the execution phase is monitored by a steering committee. Short and regular communication to all the stakeholders and personnel is also very important to support the LEAN transformation of the plant. Visualization of the LEAN key performance indicators (KPIs) through on-line display on various screens or information boards in the plant is also usually implemented, so that employees can quickly see and understand the improvements, as suggested by the TPS.  

The first case study describes the implementation of LEAN in a large-scale cell-culture facility, with three independent manufacturing lines using 3,000-L and 10,000-L bioreactors. The facility is used to produce clinical active pharmaceutical ingredient (API) material, primarily for Phase 2 and Phase 3 trials, of various biopharmaceuticals in development. Several product-to-product changeovers are performed every year in each line, resulting in a significant "loss" of total plant capacity. This facility also hosts technical development activities, in laboratories adjacent to the clinical manufacturing lines. 

Following the LEAN assessment, several specific projects were defined to increase the throughput rate, decrease throughput time, and enhance the "right first time" level. In this article, we have focused on the first two metrics. 

There are obviously numerous factors that can influence the throughput rate. For this specific facility, changeover activities were identified as having a major impact on this metric, leading to the following objectives: 

Accelerate product-to-product changeover.

Table 1 summarizes the baseline, practical vision, and achieved values for these two types of changeover activities. Under the baseline conditions, about 27% of the total yearly plant capacity was lost as a result of changeover activities. If we could achieve our practical vision target, we could reduce this capacity loss to <10%. At the end of the project, we were able to show that we could perform batch-to-batch and product-to-product changeovers in 0.55 day, corresponding to less than 3% of the yearly plant capacity. It should be noted, however, that in reality, before one can achieve such fast changeovers in a sustainable fashion, the duration of each batch must be perfectly synchronized with the timing of inoculum preparation of the next one. This requires reproducible and robust cell-culture processes. 

Table 1. Summary of the changeover acceleration project (case study 1). 

To achieve this result, each individual task performed during changeovers was evaluated, focusing on possibilities and consequences to shorten them, cancel them, or schedule them differently (details not shown). Each beneficial change was then allocated to one of the four following categories:  

low risk (regulatory, business, etc.), no investment

The combination of the first two categories was found to bring the most significant improvement in the changeover time and thus led to the changeover time of 0.55 day. In brief, the main changes were:  

radical optimization of cleaning-in-place and steaming-in-place operations

acceleration of several testing procedures

The third and fourth categories were not retained, because they brought only modest additional improvement while bearing significant risks. 

As a result of this shorter changeover time and of streamlined maintenance activities (mostly performed in "hidden time," not shown), the proportion of the yearly plant capacity available for production was raised from 68.4% to 94.9% (Figure 2). 

 can be defined in different ways depending on the scope of activities. For clinical manufacturing, we defined it as the cycle time, i.e., the total time to produce the first clinical batch, starting with technical transfer activities from the process development group. As the first LEAN milestones, the following specific projects were started:  

Minimize the lead time for the first supply of raw materials at the beginning of each campaign.

Accelerate technology transfer activities.

The discussion below focuses on the second project. 

Specifically, the goal was to reduce the generation of batch record templates and operations instructions from 14 days to 7 days with the same manpower. For this purpose, a second objective was set to reduce the total size of documents (i.e., some intermediate transfer protocols, the process description, the batch record templates) significantly. Table 2 summarizes the baseline, practical vision, and achieved values, showing that both objectives were met. 

Table 2. Summary of the technology transfer acceleration project (case study 1). 

Several improvements at different levels were implemented to achieve this result. In brief, these were:  

Manufacturing Economics Series

A step-by-step approach is essential for successful implementation.

Harvesting the Benefits of LEAN in Biopharmaceutical Manufacturing

Automated in-line dilution is an increasingly popular technology in the biopharmaceutical industry. In-line dilution is a process that can help solve capacity, financial, and quality concerns that biopharmaceutical manufacturing plants may be facing regarding process solution preparation and delivery. This technology has several applications in biopharmaceutical manufacturing, such as purification processes, chromatography systems, solvent adjustment, pH adjustment, and cleaning systems. The fundamental aspects of automated in-process dilution systems are discussed, including engineering considerations, equipment components, process materials, operation, maintenance, and quality considerations.

Automated in-line dilution is a process in which two liquid streams are brought together in a controlled fashion to meet a target diluted solution concentration. A dilution ratio of 10:1 or more often can be achieved using current equipment designs. This typically allows for concentrates of up to 10x concentration to be used as starting material. The maximum dilution ratio is limited by both equipment constraints and the properties of the concentrated solution. Much larger dilutions can be obtained by placing multiple in-line dilution processes in series. 

The equipment used for an in-line dilution process is compact and usually portable. The most basic equipment would include only one module and only perform one process step at a time, meaning that only two inlet streams are combined to make an intermediate or final product. If a second process step such as addition of another solution or adjustment of another parameter (e.g., pH) is desired, additional modules can be added to the equipment train. Multiple skids also can be placed in series to accomplish this task. An intermediate is the output of any individual in-line dilution module or skid. The intermediate is then directed to the inlet of the subsequent module to perform the next processing step. Figure 1 illustrates a process in which multiple dilutions or processing steps can be performed. 

Automation of the process allows for the final product solution to be manufactured "just in time." The small portable equipment is capable of delivering the final product at the point of use. Figure 2 shows an in-line dilution skid. 

ADVANTAGES OF AUTOMATED IN-LINE DILUTION SYSTEMS 

The use of automated in-line dilution systems provides very significant advantages to biopharmaceutical manufacturing. One of the many concerns in biopharmaceutical facilities is capacity; biopharmaceutical companies are now obtaining much higher fermentation and cell culture yields than in the past, leading to capacity shortages in downstream processing equipment. Another challenge is manufacturing at large scale. Manufacturing 10,000-L batches of process solutions in large tanks is inherently difficult. Making a 1-L solution in a laboratory can be done very precisely using analytical instruments that are calibrated at milligram sensitivity. In contrast, making a 10,000-L solution requires load cells or level probe technology with much less precision. Mixing at large scale necessitates mixing studies and process validation to ensure that the mixing process is reliable and repeatable. 

In-line dilution technology provides significant advantages compared to traditional large-scale processes because the mixing and preparation was actually being done at a small scale; compare the holdup volume of the in-line dilution skid versus a 10,000-L buffer prep tank. In addition, in-line dilution processes can incorporate feedback control with mixing to achieve highly accurate solution concentrations. 

As stated previously, a process that was originally designed for a 10,000-L batch of process solution may now require twice as much solution because of increased yields. The manufacturing process now needs to make two 10,000-L batches in the preparation vessel and transfer each batch to a 20,000-L storage vessel. Is there a 20,000-L vessel available? Is there room to install a tank this size? Figure 3 illustrates how in-line dilution can solve this large-scale problem by using a small 2,000-L tank of concentrate compared to the large 10,000-L and 20,000-L tanks needed in the previous example. 

It is not uncommon for organizations to hesitate or be uncertain about implementing new technologies. There is comfort with traditional processes where there are known failures and a general knowledge of how they are addressed. In addition to addressing new process methods, biopharmaceutical manufacturers must consider recent US Food and Drug Administration and International Conference on Harmonization (ICH) guidelines.

  These guidances urge companies to use process analytical technologies (PAT) in support of quality by design (QbD) and continuous improvement initiatives. The thought of incorporating PAT adds another level of complexity to adopting new process methodologies.  

ENGINEERING CONSIDERATONS FOR IN-LINE DILUTION

Equipment for automated in-line dilution accomplishes the following process steps: concentrated solution is combined with diluent solution, the combined liquids are mixed, solution concentrations may be adjusted depending on the system components, and the final desired solution is delivered for its intended use. Engineering considerations needed to successfully accomplish these steps are discussed in the following sections. 

There are two fundamental equipment designs for in-line dilution processes. Several variations of the basic designs are commercially available. These variations differ primarily in the method used to generate fluid flow and measure the streams. Examples of equipment designs include: 

 Fluid flow is controlled by changing the pressure of the inlet streams. For example, a beverage machine uses pressurized carbon dioxide (CO

) to mix concentrated syrup with water to produce soda with the appropriate ratio of syrup to water (Figure 4). 

 Fluid flow is controlled by measuring the mass flow rate. 

 Fluid flow is controlled by precision metering pumps. 

 Feedback control is accomplished by measuring the output of selected solution attributes. 

Parametric parameter feedback with dynamic blending.

 Feedback control is accomplished by measuring the output of select solution attributes and blending is performed in a continuously mixed blending module. 

Designs such as pressure-controlled, mass-controlled, or other mechanically based controls are not considered QbD or PAT because they do not directly identify, monitor, and control a critical process parameter (CPP). The CPP in buffer blending, for example, is the chemical make-up of the buffer (molarity and pH), not its mass, pressure, or volume. By controlling the CPP, one can eliminate the impact of out-of-specification (OOS) concentrates and the extreme difficulty and cost of trying to make nearly perfect concentrates with mass, pressure, or volume blending designs. Designs that use a blending module with PAT and feedback controls make it possible for the dilution process to decouple the variability of the final product from the variability of the concentrate.

There are three basic equipment components that comprise in-line dilution equipment. These include piping, fluid flow devices, and mixing devices. Control instrumentation includes mass flow meters and analytical instruments such as pH, conductivity, or near-infrared (NIR) instrumentation. A programmable logic controller (PLC) integrates the operation and control of all components. 

Like most equipment in the biopharmaceutical industry, in-line dilution skids are primarily constructed of 316 L stainless steel tubing connected by sanitary fittings and ethylene propylene diene monomer (EPDM) or silicone gaskets. EPDM and silicone are widely used in the pharmaceutical industry because of their favorable chemical and temperature resistance and spongy-elastic properties. Gaskets are also available in Viton, PTFE, and many other compounds. The material chosen must be compatible with the process liquids and must pass US Pharmacopeia (USP) Class VI testing. Other metals such as Hastelloy or AL6XN can be substituted for 316 L stainless steel in applications where corrosive liquids are handled; these alternate metals have different elemental ratios and different compatibility properties. Product concentrates may require special considerations for compatibility; after dilution, standard product contact materials should be acceptable. 

Fluid flow is accomplished by the following: 

 In pressure flow designs, fluid flow is accomplished by using compressed gas to pressurize the initial source of the concentrate and diluent. Since fluids flow from high to low pressure, the fluids under pressure have the ability to flow through the process because the outlet of the process has little or no pressure. 

 In mass-flow or blending module designs, rotary lobe pumps can be used. Rotary lobe pumps consist of two rotating lobes that are similar to gears. The inlet side of the pump allows liquid to flow in. The liquid is then forced through the outlet at a greater pressure than the inlet side, which results in flow (Figure 5). 

 Diaphragm metering pumps may be used in metering and blending module designs. Metering pumps move precise volumes of liquid per revolution to provide accurate flow rates. The amount of flow is changed by either changing the stroke length or by adjusting the cycle frequency. 

A centrifugal pump uses an impeller to draw in liquid at the center and force the liquid outward as it spins. The liquid gains energy (increased speed) as it is forced outward by the impeller. As the liquid exits the pump it slows down again, but now contains energy in the form of increased pressure, resulting in flow (Figure 6). 

 Static mixers are pipes with baffling inside to force a turbulent fluid flow condition. The turbulence causes the liquid to mix. 

 A blending module is a small continuous mixing chamber that is located between the inlet and outlet streams. The chamber could consist of a small tank or a large pipe (Figure 7). 

Control instrumentation monitors the control parameters for solution input to the system, and monitors parameters measuring the attributes of the diluted solution. The following are used: 

 A mass flow meter uses the Coriolis effect to measure the changes in vibration of a pipe as mass flows through it. Because there are no moving parts, this results in a reliable instrument that does not wear out and is not likely to drift out of calibration (Figure 8). 

Conductivity is impacted by the amount and type of ionic chemicals such as metals and salts in the solution. It is also impacted by solution temperature; therefore, adjustments must be made to compensate for this effect (Figure 9). 

pH is affected by temperature, but some sensors can automatically measure and compensate for temperature. Caution must be taken for product solutions that are at the extremes of the pH scale (~0 or ~14) because some sensors cannot measure at these ranges and can be damaged by these harsh solutions, which is referred to as "pH poisoning," Figure 10). 

Automated in-line dilution can help solve capacity, financial, and quality concerns that biopharmaceutical manufacturing plants may be facing.

A Closer Look at Automated In-Line Dilution

In May, I wrote about Ian Morrison's predictions that some version of healthcare reform would be better for the industry than 

 healthcare reform. A lot has happened since then, including two bills, the Senate Finance Committee's proposal led by Senator Max Baucus (D-MT), and the deal that PhRMA struck with the White House in June to contribute $80 billion over 10 years toward the cost of reform.

Even as things continue to evolve, the June deal looks pretty darn good. So good, in fact, that I have begun to think my earlier fears that healthcare reform might hurt the drug industry were rather naïve.

Through the arrangement, pharma companies would help pay for the cost of healthcare reform by providing a 50% discount for brand-name prescription drugs in the Medicare coverage gap (the so-called "doughnut hole"). In exchange, there would be no drug reimportation or price negotiations on drugs sold to people covered by Medicare. And according to various recent estimates, the additional sales that the industry may gain through healthcare reform, primarily through expanded coverage, could be significant, and far outweigh any sacrifices.

, for example, expanded coverage could bring drugmakers approximately $115 billion in additional revenue over 10 years. This number has been extrapolated from the proposed 15–20% fees on the industry's increased business—fees that are included in the industry's $80 billion contribution. (Senate Finance Committee staffers have estimated pharma's fees at $23 billion, which is 20% of $115 billion). Subtract the $80 billion in costs savings that PhRMA promised the White House, and the industry nets $35 billion.

Of course, not everyone agrees with such estimates. But it seems clear that insuring the roughly 46 million Americans who are now uninsured will increase drug sales.

It's also true that other plans to add new pharma industry fees are afloat, like the one proposed by Senator Bill Nelson (D-FL). Nelson's amendment to the Senate bill would extract about $100 million more from manufacturers over 10 years, by recovering profits the industry gained following the 2003 Medicare drug legislation. Under that legislation, older Americans who had received prescription benefits under Medicaid were shifted into the new Medicare coverage, where the government paid higher prices, because the Medicare plan did not include discounts that were part of the Medicaid plan. In the latest development, the Senate Finance Committee narrowly voted against Nelson's proposal. But even if it were to pass, it is useful to remember that such fees would basically reverse recent gains, rather than be truly new cuts into industry profits.

In a July interview on National Public Radio, PhRMA President and CEO Billy Tauzin opined that healthcare reform is not only a moral imperative but an economic one. "The studies we have indicate that if we do nothing now, this country is going to get sicker and poorer and less competitive in the world," he said.

I couldn't agree more. So although major elements of reform remain to be hammered out, I am no longer worried about whether the pharma industry, or innovation, will suffer as a result. In that same NPR interview, Tauzin himself indicated his confidence. "We'll do OK," he says.

So now I'll go back to worrying about the American people, particularly the uninsured and the underinsured, who desperately need reform. I hope we 

From the Editor

It seems clear that insuring the roughly 46 million Americans who are now uninsured will increase drug sales.

Pharma and Healthcare Reform, Part 2

A butterfly flapping its wings over Brazil can influence the direction and severity of a storm in New England. Called the "butterfly effect," mathematic models have shown that even very small changes in one area of a system can create large variations downstream.

Applied to the pharmaceutical manufacturing process, there is a "butterfly effect" that has the potential to create large savings for your operations when viewed from a more holistic perspective. It's not where you would normally look to free up working capital and reduce inventory requirements. Yet that's the effect of adopting a rapid screening method in the microbiology laboratory.

Consider this: If your company's products are free from contamination 99% or more of the time, then why are these products held up for an average of 5 to 14 days while traditional microbiological limits or sterility tests are conducted? If only the tiniest fraction, 1% of your products at most, needs to be detained for the full 5 to 14 days for further evaluation, then why subject the remaining 99% of your clean production to the same timeline? Isn't it smarter to release clean product as soon as you can determine it is, in fact, clean? What's needed is an absence or presence test, the results of which can be determined in as few as 18 hours.

Products testing negative for contamination can be released and shipped out to distribution and to customers faster, greatly streamlining your supply chain and reducing your working capital requirements. These efficiency improvements also reduce warehouse space requirements and even the amount of capital tied up in safety stock.

Recently, a Chicago-area manufacturer of a wide range of pharmaceutical and over-the-counter products found it was quickly outgrowing its 10 year-old facility. The need for warehouse space was far outstripping projections. The company faced a critical financial decision: Should they build or rent additional facilities? On examination, it became clear that an increasingly significant portion of their existing warehouse capacity was allocated to the micro-hold area. Here, formulated products were being held in sealed vats before bottling while samples from these vats were being checked for contamination. Using traditional microbiological testing methods, product samples from every batch, every day, were poured into agar plates and placed on racks where they would sit for several days, also taking up floor space. The plates could be checked at any time for organism growth, but the lot couldn't be released until there had been five days without growth.

The director of quality control was familiar with rapid methods. In fact, the company had a system in place, but it was capable of running only a few dozen tests per day. When pushed harder, it invariably broke down and needed service. Something had to change. The company selected a new, high-throughput screening system using adenylate kinase, which is used to enhance an adenosine triphosphate (ATP) bioluminescent reaction. The quality department now has the capacity to screen up to 120 incubated samples per hour. Lots are routinely tested; those absent contamination are packaged and released.

For this company, a capital investment of approximately $100,000 returned a five-year net present value of nearly a million dollars. This relatively small flap of a butterfly's wings in the microbiology laboratory to a new, more efficient and reliable system, made a large difference downstream, enabling the company to reclaim significant warehouse space and extending the life of its facility for at least another 10 years.

 is a vice president at Celsis International, Chicago, IL, 312.476.1200, 

Final Word

Small changes can have a big effect further downstream in your manufacturing processes.

Apply the "Butterfly Effect" to Streamline Your Supply Chain

The contract manufacturing world was roiled in August when Lonza (Basel, Switzerland), a market leader in manufacturing large-and small-molecule active pharmaceutical ingredients (APIs), announced an offer to acquire Patheon (Research Triangle Park, NC), the second largest contract manufacturer of drug products. Lonza offered to buy shares of Patheon for $3.55 per share, valuing the company at about $700 million (market capitalization plus debt).

Most industry participants were stunned by the audacity of the move: Lonza was wading into an intense battle for control of Patheon already being waged by the private equity firm JLL Partners (New York, NY) and a committee of independent directors representing non-JLL shareholders. JLL had been Patheon's white knight two years ago when it needed to restructure its finances, providing $150,000 in exchange for preferred shares of the company. Earlier this year, JLL offered $2 for the outstanding shares of Patheon, and now owns 57% of the company (including the conversion of its preferred shares to regular shares). That tender offer triggered a series of public relations and courtroom skirmishes with Patheon's independent directors over the value of the company. The independent directors, believing the company is worth more than $2 per share, have welcomed Lonza's offer.

Beyond that initial reaction, however, the industry is questioning Lonza's stated rationale for wanting to acquire Patheon. Lonza has been quite clear for some time about its interest in adding formulation development and dose manufacturing to its service mix. The aim is to make Lonza into a turnkey supply chain solution for bio/pharma companies, able to deliver both the API and finished drug product in the major pharmaceutical markets. As recently as July, Lonza's CEO Stefan Borgas stated that the company had $1 billion available to spend on acquisitions, and that dose manufacturing was a principal target area for acquisitions.

To the degree that strategy makes sense, acquiring Patheon would enable Lonza to implement it on a large scale. The contract manufacturing organization (CMO) already has the following advantages: 

Patheon is the second-largest dose manufacturer, with $570 million in revenues from manufacturing commercial products.

It has manufacturing operations in five countries on two continents (three facilities in the US, three in Canada, and four in Europe including two in Italy, one in France, and one in the UK) spanning most major dose forms.

It has one of the largest drug development operations, with revenues of $130 million.

It has successfully completed a painful restructuring process that has driven a lot of cost out of its operations and positioned it for significant revenue and profit growth.

However, Patheon's attractiveness as an acquisition candidate aside, there are still a lot of questions about the viability of the one-stop manufacturer strategy. Among the major objections that might be raised about the strategy are the following:  

Lonza's bid for Patheon makes the contract manufacturing industry re-examine the one-stop manufacturer model.

BioPharm International-10-01-2009