The 5-Step Chemical Batch Review Checklist That Caught What Our Site Audit Missed

Over the past four years, I've reviewed somewhere around 200 chemical and biological products a year before they ship. APIs, cell culture media, specialty biocides, and commodity chemicals like hydrochloric acid. The judgment call—"is this batch ready?"—looks straightforward from the outside. The reality is that a release decision pulls in documentation, history, and spec data that most buyers never see.

This checklist is for quality managers, procurement leads, and contract manufacturers who approve chemical suppliers or take delivery of critical materials. If you're in pharma, biotech, or specialty chemicals, these are the five checks I actually run—not the glossy version that shows up in sales decks. It's a working framework, not a theory.

Step 1: Trace the actual manufacturer—not the label

People assume the brand name on the CoA (Certificate of Analysis) tells you who made the material. Often it doesn't. The brand on the label may be decades old while manufacturing has moved to a different site under a different quality system entirely.

Here's a real example. Arch Chemicals built a strong reputation in biocides and specialty chemicals. When Lonza acquired Arch Chemicals in 2011, the product lines kept their legacy names for years, but production shifted under Lonza's quality systems. If you sourced "Arch" material into 2024, you were getting Lonza-manufactured product. I can't tell you how many approved supplier lists never updated to reflect that.

For the chemical hydrochloric acid, the same traceability principle applies. HCl seems like a simple commodity—"it's just acid." But the chloride source, the distillation method, and the packaging liner all affect trace impurities. We've tested HCl from two vendors with identical specs on paper and different iron content in practice. The difference traced back to a different production route.

Check for:

• The legal manufacturer on the CoA matches the site that actually produced the lot
• Legacy brand names (like Arch) are reconciled with current ownership
• Manufacturing site addresses are consistent with regulatory filings

Step 2: Demand documentation beyond the CoA

A Certificate of Analysis is a starting point, not proof. It tells you what the supplier measured on one batch. It doesn't tell you whether the test methods were validated, how the batch has behaved under stability conditions, or whether the material has really been consistent over time.

In our Q1 2024 quality audit, we flagged a vendor whose CoA looked immaculate but whose test method was an in-house protocol with zero validation data. The product may have been perfectly fine. "May have been" doesn't pass a regulatory inspection.

This is one area where the industry has genuinely changed. What satisfied quality audits in 2020—a one-page CoA with five basic parameters—is no longer enough. Pharma companies, including emerging players like Celon Pharma, now request complete documentation packages: method validation summaries, stability data, impurity profiles, and batch record traceability. I do not see that trend reversing.

Check for:

• CoA showing batch-specific results, not "conforms to specification" boilerplate
• Analytical methods referenced by ID with validation status
• Stability data aligned with your intended shelf life and storage conditions

Step 3: Scrutinize the formulation spec—not just the label

"RPMI-1640" on a label suggests a standard product with a standard spec. If you've worked in cell culture, you know that's not the full picture. Lonza media formulation work is a useful example. We produce RPMI-1640 and DMEM under quality systems that control for osmolality, pH after reconstitution, endotoxin, and bioburden—not just the amino acid and vitamin profile printed on the label.

Those CQAs (critical quality attributes) are what determine whether a media batch performs. I remember a 2022 batch that passed every release specification but had a slight color shift. We debated it internally and shipped it anyway. Our customer's cells noticed within 48 hours. Root cause: a lot of L-glutamine that was technically within spec but sitting at the edge of the acceptable range. If I remember correctly, the vendor's own data had shown a drift trend for three lots prior—we just hadn't asked to see it.

We tightened the spec and changed how we assess incoming raw material lots. Now we ask not just "does it meet spec?" but "where in the range does it fall?" It's a small adjustment with outsized impact.

Check for:

• Full formulation documentation, including raw material grades
• CQAs defined with tight ranges, not broad tolerances
• Appearance or color criteria included in release specifications

Step 4: Review the toxicological and environmental data yourself

This is the step most people skip because it feels like "a regulatory person's job." In practice, that often means nobody reviews it with real care.

Let's take a common question: what does atrazine do to humans? It's not a simple answer. The EPA set a maximum contaminant level of 3 parts per billion in drinking water; the European Union banned atrazine in 2004. The published research on endocrine effects is substantial and at times contradictory. That regulatory divergence is exactly why you cannot rely on assumption or a decade-old safety data sheet. The data evolves, and your quality review should evolve with it.

Whether you're working with atrazine, a chemical intermediate, or the chemical hydrochloric acid, the toxicological and ecotoxicity data should be current, sourced, and reviewed by someone on your team—not just filed away.

Check for:

• A current SDS (not the copy from three years ago)
• REACH registration status and exposure scenarios, where applicable
• Human health and ecotoxicity endpoints with study references

Step 5: Verify with your own tests—at least on first batches

Every spreadsheet I ran pointed to one of our suppliers being dependable. Pricing was competitive. CoAs were consistent. Lead times were reliable. And still, something felt off—specifically, they were slow to answer basic questions, and the answers felt rehearsed.

I trusted the data and pushed the hesitation aside. Later, when we ran independent identity testing, we found that two consecutive lots had near-identical CoAs but measurably different solvent profiles. The supplier had changed their cleaning procedure and decided it didn't matter. It did matter.

Now we reserve reference samples from every batch that goes into GMP production—or rather, from representative lots at a frequency that's defensible in an audit. We test the parameters most likely to drift: appearance, identity, and pH. The cost is minor. The insight is not.

Check for:

• A defined first-batch test plan before the material arrives at your dock
• Reference samples retained for at least the product's shelf life
• Re-verification of identity whenever a supplier changes process or site

Common Mistakes I See Routinely

Not every supplier problem shows up as a failed specification. The patterns that cost the most are the quiet ones:

1. Treating the CoA as evidence. A CoA reports measurements. It doesn't prove the methods behind those measurements were valid, or that the sample truly represents the batch. If you can't explain the test method, you don't understand your own supply.

2. Trusting legacy brand names. The Arch Chemicals Lonza transition taught our industry a broader lesson: names persist, but ownership, sites, and quality systems change. Reconcile your supplier list annually.

3. Assuming "within spec" means "safe." A parameter at the edge of the range can still cause problems in your process. Watch for drift, not just outliers.

4. Delegating toxicological review to a department that may not exist. If your company has a regulatory affairs team, great. If it doesn't—and many don't—quality owns it. Read the SDS.

Final Thoughts

The fundamentals—verification, documentation, consistency—haven't changed. What's changed is the complexity of chemical supply chains and the expectations placed on them. In 2020, a CoA satisfied most audits. In 2025, it's the starting baseline, and the bar keeps rising.

This checklist works in our context at Lonza, where quality teams review a broad mix of APIs, media formulations, and specialty chemicals. If your operation is smaller or more niche, you may not need the full documentation depth on every order. But the structure—trace, document, inspect, review, verify—translates. Or rather, it translates if you actually follow it, which is the difference between a quality system and a binder of paperwork.