evolutions in direct compression

52 Pharmaceutical Technology APRIL 2005 www.pharmtech.com

SPECIAL REPORT

very solid-dose formulation team starts out aim-ing for a direct-compression tablet. It’s the moststraightforward, easiest to control, and least ex-pensive manufacturing option. Direct compres-sion, of course, uses two primary process steps,mixing and compression, to produce the finished

tablet. Wet granulation adds five more operations: wet gran-ulation, wet milling, drying, milling, and mixing again. Eachadditional step requires equipment, operators, space, power,and expense…and every step lowers the yield and increasesthe risk of out-of-specification product.

So the most attractive solution almost always is the fa-miliar, undramatic standby, direct compression.

“The basic tablet press has been fill, compress, andeject since about 1875,” says a veteran equipment

marketer. “Ever since then, it’s been a matter ofhow do you control it better and now how do you

clean it better.”Yet, at last November’s annual meeting

of the American Association ofPharmaceutical Scientists (AAPS),a crowd of hundreds overflowedthe Baltimore Convention Center’slargest meeting room for the ses-sion that promised “New Devel-

opments in Direct Compression.”Why the standing-room-only atten-

tion right now? “It’s not ‘right now’,” says consultant Metin Çelik, PhD,

(Belle Mead, NJ, www.pt-int.com). “Direct compression isalways there.”

Larry L. Augsburger, PhD, professor of pharmaceutics atthe University of Maryland (Baltimore, MD, www.umary-land.edu), studies solid oral dosages. “You probably have alot of people hoping to learn something,” he said. “They’retrying to expand to a broader range of drug loading. And

you might have a lot of people who don’t have a basicgrounding in the pharmaceutical sciences.”

Direct compression is easier and more economical…whenit works. Or, rather, when it works within the allowable time.

Time vs. moneyTime pressures vary over the drug-making continuum, fromtime-sensitive, high-margin new prescription drugs at oneend to durable, low-margin over-the-counter and nutraceu-tical formulations at the other.

For a new prescription product, every day or week lostin formulation wastes irretrievable patent life. Speed, con-sistency, and reproducibility are the watchwords. So formu-lators spend relatively little time trying to attain a success-ful direct compression before moving on to easier-to-developwet-granulation techniques.

“Every time we have a development project, we start withdirect compression,” says Gary Bubb, vice-president of SMI(Lebanon, NJ, www.smitmc.com), a vendor of tablet pressesand controls. In the end, though, he estimates that perhaps80% of NDA projects wind up as wet granulations.“The de-cision is primarily driven by timelines rather than by cost. IfI had a tight deadline, for example, I would do a wet granu-lation. It’s the most forgiving and most likely to succeed. Imight not even explore anything else, unless I’ve got the re-sources so that I can explore things in parallel,”Bubb explains.

In generic drug formulations, time pressure is lower andcost pressure is higher. The products have lower margins andthe processes will generally run longer. Reducing costs via di-rect compression can have a bigger proportionate impact ontotal profits. In this environment, says Ruey-Ching (Richard)Hwang, PhD, a senior director of pharmaceutical sciences atPfizer Global Research and Development (Kalamazoo, MI,www.pfizer.com), “more than 50% of the generic formula-tions I’ve worked on use direct compression. About 30 to 40%wind up using wet granulation. And, 10% or less use roller

E

KO

RS

CH

AM

ER

ICA

Evolutions in DirectCompressionDouglas McCormick

Evolutions in DirectCompression

The principles of directcompression haven’t changed inmore than a century. So why canthe topic still pack a ballroom?


compaction.”Hwang calls this last technique“underutilized.”

For the most part, though, formulatorscan only look with envy on the direct-compression success reportedly achievedby Johnson & Johnson’s former develop-ment director, Paul E. Wray.“He spent hisentire life doing formulation develop-ment,” says one equipment maker whoworked with him,“and in his entire careernever had one product that was not directcompression: If you fight long enough andwork hard enough and don’t take the lazyway out, you can make it work.”

Signposts for successful compressionWhat are the barriers to a successful di-rect-compression formulation and whatcan go wrong after the process goes intoproduction?

At AAPS, Pfizer research fellow BrunoC. Hancock cautioned the audience withan outline of Things That Can Go Wrongwith Direct Compression (see Table I) .Hancock emphasizes how important it isto understand the properties of theblended active pharmaceutical ingredient(API) and excipients, starting with parti-cle shape and size distribution, frictionand adhesion profiles, true and bulk den-sities, compaction parameters, and com-pacted mechanical properties.

Four factors are critical in a manufac-turable direct-compression formulation,Hwang says: blend uniformity, blendflowability (especially at the rates de-

manded by high-speed presses), compress-ibility, and lubricity. Of these, the mostimportant is flowability, though the prop-erty is disturbingly difficult to quantify. Itcan be measured as an angle of repose oreffective internal friction (a small angle isgood; a large angle is bad) or by flow ratesand performance in a test funnel or hop-per (mass flow, good; rat-holing, bad). Theflow-rate tests, however, depend on thesize, shape, and even the material of thetest vessel in a way that would make a the-oretical physicist cringe.

Workers in the field often hesitate to

cite specific cut-off values for these keyproperties; Hancock does not. His AAPSpresentation culminated by listing, andquantifying, the hallmarks of an ideal di-rect-compression material (see Table II).

Without well-established, well-quanti-fied indices to distinguish good blendsfrom bad, formulators must fall back onexperiment. Armed with sufficient APIand excipient materials and a design-of-experiments approach, Hwang says, a for-mulator should know whether a direct-compression formulation is feasible aftertwo sets of studies, consisting of 10–20 ex-periments each. (The whole programmight take a month or two.)

Understanding excipients…better Ultimately, expanding the useful range ofdirect compression requires expandingour understanding of how the APIs andexcipients behave as bulk powders andunder compression, transforming the cur-rent qualitative and subjective judgmentsof “good” into quantified, predictive parameters.

“There’s a growing body of informa-tion cataloging the physical properties ofdirectly compressible materials, but theinformation is scattered,” says Augsburger.“For example there have been some recentpapers on using multiple methods or in-dices to characterize the flowability of di-rect-compression fillers. Instead of saying,‘This is a very compressible excipient,’ you

SPECIAL REPORT

Table I: Direct compression pitfalls.Problem Effect Key factors

Poor blenduniformity

Inconsistent tablets

Particle size and shape, mixer design, poor blending design, segregationduring handling

Erratic powder flow

Tablet weight variation,press adjustments, lowpress speeds

Particle size and shape, equipmentdesign, high internal friction, externalfriction with vessel walls

Low blenddensity

Difficulty achievingtablet weight, pressfeeding problems

Density, true and bulk

Poorcompressibility

High punch forces, soft tablets

Particle size and shape, powder ductility,tooling design, press set up

Mechanicalweakness

Friability, logo picking,erosion during coating

Particle size and shape, tensile strength,tooling design, press set up

From BC Hancock, “Identifying Candidates for Direct Compression Using Material-SparingFormulation Tools,” presented at AAPS, November 2004.

Table II: Properties of an ideal direct-compression material.Property Parameter Target value

Particle size andshape

D 4,3 (mean volume diameter)D 10 (10th percentile diameter)D 90 (90th percentile diameter)Aspect ratio

.80 mm

.30 mm,1000 mm,1.5

Blend uniformity Blend potency ,2% RSD*

Powder flow Effective angle of internal friction ,41˚

Powder densityTrueBulk

.0.5 g/ml1.0–2.5 g/ml

Tabletingperformance

Dwell time sensitivityPrecompression forceCompression stress (at ;0.85 solid fraction)

LowLow20–125 MPa

Compactmechanicalproperties

Tensile strength (at ;0.85 solid fraction)Brittle fracture index (at ;0.85 solid fraction)Indentation hardness (at ;0.85 solid fraction)

.1.0 MPa,0.275–250 MPa

* RSD is relative standard deviation From BC Hancock, “Identifying Candidates for Direct Compression Using Material-SparingFormulation Tools,” presented at AAPS, November 2004.


might be able to say, ‘It has a bondingindex of such and such and it has a flowa-bility index of such and such,’ giving a bet-ter comparison for choosing and select-ing. Even the concept of compactability isnot well understood right now….”

Hwang notes that several industry research groups have mined their ingredi-ent-characteristic and formulation-per-formance databases, trying to correlate stan-dard ingredient-property measurements(including various indices of flowability)with later scale-up and manufacturing be-havior. The resulting measures can give theformulator a ballpark expectation of suc-cess,“but there’s never a one-to-one corre-lation between index A and final behaviorB,” says Hwang.

Hancock’s group has had some successusing shear cells to quantify flow charac-teristics. It’s an old and sometimes cantan-kerous technology, he admits, but it yieldsmore consistent and comparable data.Hancock notes that his laboratory buildsabout half of the research equipment theyuse, using what he calls a “science of

scaling down.”Rather than begin with com-mercial analytical equipment and thenscaling the formulation up, the Groton, CT,research group starts with manufacturingequipment and then builds laboratory-scaledevices that mirror its behavior.

Post-approval traumaThe formulation challenges don’t end whenthe process is validated, up, and running.

Though the very thought of reformu-lating an approved product is anathemato most process managers, sometimes “re-formulation is the easier way to go,” in thewords of one excipient supplier. “In theshort term, it’s expensive, especially if thereis some bio work required to revalidatethe process.” Still, reformulation may bethe best, most economical, or indeed theonly way to resuscitate a product that hasgotten into regulatory trouble because ofpoor tablet consistency or physical defects.

Some tableting problems, says an equip-ment manufacturer, “are the result of justso much international sourcing of ingre-dients. Today the purchasing folks are

under so much pressure that they buy thisbatch from here and that batch from there.Even though they pass quality control,their characteristics are not exactly thesame. In production, you can’t get thebatches to turn out the same no matterwhat you do to them.”

When a tableting process breaks down,the consequences can be dire. In the mostrecent case, federal marshals seized allstocks of two GlaxoSmithKline productsfor what appear to be tableting errors. (Seenews item on page 21.)

Tableting in evolution“Though there have not been any impor-

SPECIAL REPORT

Compressibility

High Low

High

LowFlo

wab

ility

Wetgranulation

Drygranulation

Direct compression

Powder compressibility and flowabilty are thekey factors in choosing a tableting method.


tant developments in direct compression forpushing ten years,” says Augsburger, “therehave been some innovations in the equip-ment. The evolutionary process goes on, butthe basic technology depends on the excip-ients and whatever API you’re using.”

The important concentration in recentyears has been in better machine manage-ment and better integration with the restof the production process, says Dale Na-toli, of tool-and-die maker Natoli Engi-neering (St. Charles, MO, www.natoli.com)“We’re not talking innovation here. We’renot talking new, “ he says. “We’re talkingproper management.”

Process analytical technologies (PAT). Ifthere is indeed a growing interest in thefundamentals of direct compression, theopportunities of process analytical tech-nology may be responsible. “With fewersteps, and fewer process variables, theprocess should be easier to model, moni-tor, and control,” says Çelik.

Augsburger concurs.“It may sound likebandwagon stuff,” he says, “but I do thinkthat some of the ideas being stimulated by

PAT initiative—‘Hey, we’d better go backand understand our processes or formula-tions a whole lot better than we do now’—could have a payoff that’s significant. Sonow we have people scrambling to do basicpharmaceutics where they never did it be-fore. It’s pretty exciting.”

Right now, tablet-press data-collectionfocuses on the product (controlling tablet-weight, for example) rather than on themachine (controlling operating temper-atures or providing predictive mainte-nance). Even so, many tablet presses al-ready provide more process informationthan the operators now use.

Data acquisition systems—originalequipment or add-on—are available forold and new systems. They are capable ofproducing reams of data, like detailedcurves of compression force vs. time foreach tool, ejection forces, and strippingforce. (Ejection force pops the tablet outof the die. Stripping force pushes it off thelower punch face.)

Unused information. In general, say mostexperts, the industry focuses on a single

data stream: peak compression force. Andmost processes monitor just a single aspectof that: peak compression force variation.Low variation indicates consistent tabletweights; larger variations indicate that theamount of material filled into the die is in-consistent, so that tablet weights will vary.

The other data are ignored.“Unfortunately the industry doesn’t

know what to do with this information,”says a press-maker. “Great, we’ve spent$60,000 or $80,000 or $100,000 and whatdo we do with it now? They have a lot ofinformation they could use, and it’s notjust looking for peaks.”

The compression-force curve alonecontains a good deal of data that isn’t gen-erally used. Dwell time is critical to ensur-ing proper compression: make it too short,and tablets are likely to cap (separate atthe top layer) or laminate (separate at aninternal layer). Roughness in the forcecurve can warn of undue friction in thepunch or die, and indication of deforma-tion and impending failure.

Ejection and stripping force readings

SPECIAL REPORT


can be useful diagnostics, says Natoli: theseforces generally increase when heat—generated by compression and friction—builds up in the tooling.

The heating problem deserves more at-tention, in Natoli’s view. “The biggestproblem in this industry is picking [smallparticles adhering to the tool faces, result-ing in small defects in engraved logos andedges] and sticking [heavier materialbuildup on the tooling].” Right now, Na-toli says, researchers don’t know exactlyhow much of the heat buildup inside thepunch is caused by friction and how muchis caused by the thermodynamics of com-pression. Powders come together andabrade across the punch face, he says, cre-ating a superficial heat much higher onthe tool surface than it is inside the toolor machine. Over time, operating temper-ature rises, which can make product stickto tool faces and die walls. The high tem-perature can even burn the materials,causing black specks and streaking of thetablet’s side wall.

Keeping it cool. “We want the press to

run as cool as possible,” Natoli says.“If wecan reduce the amount of heat we’re cre-ating, then we’re going to eliminate a lotof tableting problems: sometimes …thesurface of the tool gets warm and the ma-terial melts on the surface of the punch.And these data acquisition systems can’tdetect that directly. They catch it when thetablet-stripping force increases, and thenwe realize that we’ve started to film. Oncewe start to film, then we start to stick andpick. We want to eliminate stick and pick,and so we want to monitor tablet takeoffand ejection forces.”

There are limits, of course, to how use-ful these data can be in practice. Natoli ap-plies a practical standard: “If you have aproblem product that tends to bind andstick and [cause] premature tool wear,then I would say it should be monitoredconsistently.” If developers have producedan efficient combination of blend, tool-ing, and press set-up that runs fairly well,then spot monitoring is sufficient.

Containment, wash-in-place, and layering.The general manufacturing trends toward

process containment and washing-in-placehave commanded much of the press-makers’ R&D attention over the past fewyears. “Containment and wash-in-place(WIP) have been a big deal,” says one in-dustry observer.“All of the [upstream] op-erations have been contained, so now we’removing to the end of the line, the final op-eration is being contained, with WIP aswell.”

The move to containment goes hand inhand with the increasing prevalence ofhigh-potency active ingredients and theincreasing focus on operator protection.

Containment clearly makes the plantcleaner and easier to manage. So does WIPequipment, which requires only minoroutward modifications in the design. ForWIP, “materials and construction have tobe compatible with the cleaning agents.Windows have to be sealed properly. Andthere has to be some way to remove thewater that makes sense. It doesn’t look dra-matically different to a casual observer,though,” the observer says.

He also notes an increasing interest in

SPECIAL REPORT


multilayer tablets. “They’ve been around for many years andthere’s nothing new in terms of feasibility,” he says, “but sud-denly everybody seems to want to make them and make themfor real.”

Tooling trendsIt’s probably natural that Natoli, a tablet-tooling maker, feelsthat tooling often fails to get the process-management atten-tion it deserves. Then again, the punch and die touch the tablet,and the tablet winds up in the customer’s hand bearing the im-prints of that operation. Proper tool management is thus vitalto consumer acceptance…as well as to the in-specification acceptance rate.

The evolution in tooling has centered on productivity anddurability: multiple tool tips, new tool materials, and rotatingpunch heads.

Multiple tool tips.Multiple-tip tools produce more than one tabletper station per pass. With 10–99 stations and one or two passesper station per revolution, multiple tool tips with as many as fourtablets per tool greatly increase the tablet production rate. TheseEuropean-style tools, Natoli notes, have been common abroad forseveral years, but have been used in the United States only for thepast five years. They do increase demands on the press: the totalcompression force increases as a multiple of the number of tabletsin the die, and ejection and stripping forces increase as well.

Multiplied compression forces, faster press speeds, and mul-tiple compression cycles all increase the wear on the tool head(i.e., the portion that rides in the guide tracks and contacts thetablet press’s compression rollers). Natoli’s company has re-sponded by allowing the tool heads to rotate as they travelaround the press, thereby spreading the wear over the wholetool head and lengthening head life.

Tool materials. For decades, machine shops have made phar-maceutical dies and punches from standard tool steels. Today’scompression tooling makers have just begun to explore pow-dered-metal and composite (e.g., high-density plastics) mate-rials. “These are fairly new, so there’s a whole lot to be learnedat this point; we’re just in the process of trying to reformulatethese materials to tailor them for pharmaceutical compressions.”Natoli says.

Composite materials are available on what Natoli calls strongertablet configurations (i.e., those with smaller cup depths). De-signers go to deep cupping for esthetics—it makes the tabletappear smaller and easier to swallow. “It comes down to con-sumer acceptance,” Natoli says.

The deeper the cup on a tool, the weaker the cup edge. Thisincreases the potential for distortion during compression, lead-ing to brittle fracture along the punch edges. Flatter tablets havesharper corners, which give a stronger tooling configuration,but may give an appearance of being uncomfortable to swal-low. The sharper corners of a flat tablet may increase chancesof chipping: “The industry certainly believes that,” says Natoli,“but I don’t think this is always true.”

The important thing, he says, is communication between thetool manufacturer and the tablet maker,“and they’re not doingit as well as they should to provide a configuration that’s accept-able to marketing, consumers, and the tool manufacturer.” PT

evolutions in direct compression

Documents