

An enzymatic route can cut the waste of a pharmaceutical process several-fold. The number that proves it, however, is built before the comparison, not after: if the boundaries of the mass balance shift between the chemical and the enzymatic route, the result says more about the accounting than about the chemistry.
Whoever has to decide whether to replace a step needs a figure that holds up in front of production, purchasing and the sustainability report. In our biocatalytic development work we set it up this way: a mass balance per unit operation on the existing route and on the enzymatic one, the same boundaries for both, and the techno-economic assessment alongside.
Sheldon’s E-factor is kg of waste per kg of product. Process Mass Intensity (PMI), adopted by the ACS GCI Pharmaceutical Roundtable, puts everything that goes in on top: reagents, reaction and purification solvents, catalysts, water. Count the same materials and PMI = E + 1. The ideal is 1 for PMI and 0 for the E-factor.
Pharmaceuticals start from an E-factor of 25 to over 100 kg/kg, against less than 0.1 in oil refining. The reason is nearly always the same: solvents make up to 80-90% of the non-aqueous mass used to manufacture an API.
Biocatalysis often runs in aqueous buffer, and here the metric can turn against the enzyme. The E-factor was born without water; in pharma the trend today is to include it. The enzymatic process for the atorvastatin intermediate has an E-factor of 5.8 without process water and 18 with it. Both are correct. Comparing one with the other is not.
The same goes for solvents. Sheldon separates the simple E-factor (sEF, no solvents or water), useful for route scouting, from the complete one (cEF, all materials, no recycling), useful for waste-stream analysis; the industrial figure falls in between. For sildenafil the two values, on the same process, are 3.9 and 50.3. A comparison with an sEF on one side and a cEF on the other moves the verdict by an order of magnitude.
The most common objection is a fair one: the enzyme has to be expressed and purified, and fermentation uses water, media and energy. Becker, Lütz and Rosenthal measured the lab-scale production of 10 g of a recombinant enzyme: about 37,800 g of waste per gram of enzyme and 603 kWh of electricity. Their conclusion is that this contribution should not be neglected.
That burden, though, is divided by the grams of product each gram of enzyme converts. In the atorvastatin process the enzymes and the cofactor account for less than 1% of the waste: at the stated loadings, about 280 g of substrate per gram of ketoreductase. The levers are enzyme loading, residual activity at the end of the reaction, and reuse. And a lab figure is not an industrial one: the same authors estimate that a ten-fold higher cell density would bring energy down to about 200 kWh.
Again in the atorvastatin case, 51% of the waste comes from solvent losses, 25% from gluconate and 22% from salts. Once the enzyme is in place, what remains sits in the work-up: extractions, solvent swaps, crystallisations.
The advantage of an enzymatic step is therefore often measured in the steps around it: one protection fewer, no metal to remove, a crystallisation saved. That is why the comparison is made on the whole route, which is the ground of chemo-enzymatic processes. The reference case agrees: the sitagliptin transaminase process cut total waste by 19%, and the real gain was also in yield and in removing rhodium.
PMI and E-factor weigh kilograms, not hazard or energy: a kilogram of dichloromethane and a kilogram of brine count the same. The numbers should be read with a solvent selection guide and, when the choice matters, with a life cycle assessment. They remain the first filter, and the cheapest: they can be calculated from lab data in the first weeks, before scale-up.
In our projects the count follows three rules:
They say the same thing from two sides: PMI = E + 1 when the same materials are counted. PMI is the pharma standard and is easier to audit, because it is measured on inputs; the E-factor is more intuitive when the discussion is about waste.
Yes, as long as it is declared and applied to both routes. An enzymatic process in aqueous buffer compared without water against one in organic solvent, or the other way round, gives a result that does not survive review.
Yes. On the first experiments the simple E-factor is enough to discard the worst routes. The complete PMI comes once reaction conditions and work-up are fixed.
Technical content and news from our laboratories, sent when there is something worth saying.