Laboratory Controls Cited as Major Concern in Field Inspections

Lab controls were highlighted in FDA official’s remarks on key inspection issues.

An FDA field official last month provided insight into the agency's approach to inspecting labora­tories and lab-related computer systems. Such systems are coming under closer scrutiny due to FDA's growing concerns about data integrity ("The Gold Sheet" April 2007).

Inadequate laboratory controls are a major com­pliance problem, said Nancy Rolli, pre-approval manager with FDA's New Jersey District, in remarks focusing on her experiences in inspecting firms' laboratories and computer systems in New Jersey.

Other significant lab issues were insufficient testing and release methods for samples, she said at a July 23 conference in New Brunswick, N.J., sponsored by Pharma Conference and the University of Rhode Island.

Rolli noted that violations of laboratory controls found in Section 211.160 "has always been one of the top 10 observations…. The biggest issue we have is lack of documentation of lab controls."

Inadequate lab controls were cited in five of the 20 GMP-related warning letters FDA issued in fiscal year 2006 ("The Gold Sheet" April 2007).

Having an improperly trained staff falls under the remit of inadequate lab controls, said Rolli. "I might just pick one or two analysts and go through their records and see what is there. I want to see GMP training. You may be able to test a product but you have to know what to do if it fails. This should be in their GMP training. You want them to be trained in your procedures."

And Rolli noted that if firms do provide this training, it should be documented in writing.

"You have to document analyst training. Some firms can say they provided this analyst training but it has to be written down. And it needs to be accessible during an inspection."

Accessibility of SOPs Can Be a Problem

Inadequate lab controls also show up in inspections where SOPs either are not followed or are not accessible to lab personnel. Rolli said that "I can't tell you how many times I ask a simple question such as what do you do when a sample comes in and I can get two or three different answers on how it gets in, how it gets assigned, and how it gets tracked."

Rolli further noted that at one facility, SOPs were only available electronically, but that not everyone had access to the electronic copy.

"This company said we're not keeping any SOPs on the floor. The only problem was not everybody had training. It was very embarrassing for the firm when the people could not access the SOPs. They should be available to everyone."

Rolli further recommended that SOPs should include information on: who should make changes if there are problems; procedures for equipment calibration and methods for change control; procedures for investigating deviations; and details of the stability program.

Rolli noted that "it's really important that you stay on top of the stability program. FDA investigators are looking at that. We're going to be looking at whether you have the chambers set up for the stability samples. How often are you pulling the samples? I find a lot of time with samples is that everyone is interested in getting products out the door, so if anything suffers, it's always the stability samples.… They're behind six months. They're testing it after expiration.… It's important to retain samples."

Inadequate Equipment Calibration Seen

Another problem area related to insufficient lab controls is inadequate equipment calibration. This can include poor oversight of outside vendors if the calibration work is contracted out.

Rolli noted that in a recent inspection, a firm that used an outside vendor never looked at the results of the calibration. "They took the calibration results and threw them into a pile." She asked to see the results and found some problems; they showed that the equipment was out of tolerance, so the firm had to go back and recalibrate.

"If you're having this done by an outside vendor, you have to review it and make sure that it's what you want," she said. "You have to have a strong calibration program. If you find out that your equipment is out of tolerance, that's a problem. Don't rely on an outside company to do the work for you."

Rolli recommended that to avoid these and other equipment problems, firms should use logbooks, which should include the date the equipment was used, the product and lot number tested, the name of the analyst who entered the information, when the equipment was last cleaned, when it was last calibrated or recalibrated and the date it was last repaired.

Violations of Testing and Release Methods Noted

After inadequate lab controls, the second major GMP problem is violation of Section 211.165 governing testing and release standards.

This section specifies that any sampling and testing plans "shall be described in written procedures that shall include the method of sampling and the number of units per batch to be tested; such written procedures shall be followed."

Section 211.165 was cited in six of the 20 GMP-related warning letters FDA issued in fiscal year 2006 ("The Gold Sheet" April 2007).

One problem related to testing and release standards is a lack of specificity on which testing method is being used. Rolli said that "people try to tell me they use the modified USP. That's not the same thing as USP. It's either USP or it's not the USP."

Rolli further observed that samples should be tested and validated for U.S. approval in the U.S., not overseas. In one inspection late last year, an investigator told Rolli that the company was flying one of their analysts to India, where the drug was manufactured, to validate their test sample.

Rolli noted that "instead of taking the sample from the lab in India and shipping it here to do their criteria and acceptance sample, they flew the analyst to India to make sure the analyst could test the product in India. The analyst said, 'do you know how expensive it was for us to do this for every batch?' Transfer the product, not the analyst."

Rolli said other inspections have revealed problems with a firm's handling of samples. "It sounds like a basic thing but you would be so surprised as to how when you go on inspections and ask how a sample is received and how it is logged in. There needs to be accountability. There need to be written proce­dures for samples. There must be a written policy for procedures and sample plans and a sample collection. These are little things but you would be surprised that there are some 483s on this."

Firms, said Rolli, should have a written policy or procedures for samples that should include the sample size requirement and the sample plan, a description of the sampling containers and the storage condition for the sample, and the sampling techniques for components, water, particulates and microbes.

"GMPs require that you have to have the sampling because inspectors will be looking at manufacturing procedure, but there are no requirements on how to pull a sample. A sampling documentation should go along with your validation. There has to be documentation of validation sampling."

Insufficient Laboratory Records Also Noted

Following inadequate testing methods, the third major problem area Rolli noted is insufficient laboratory records. One firm in Rolli's jurisdiction received a 483 observation for violating Section 211.194 governing laboratory records.

Violations of Section 211.194 were cited in five of the 20 GMP-related warning letters FDA issued in fiscal year 2006 ("The Gold Sheet" April 2007).

Section 211.194 requires that laboratory records include complete data derived from all tests necessary to assure compliance with established specifications. The lab records should include a statement of each method used, a statement of the sample weight, and a record of all test calculations.

The firm that received the 483 had no documentation showing problems with the sample collection.

"They were getting some failures from their in-process samples. They did three validation batches and all three failed. So we went through the batch records…. When we looked at the batch records, they did not say anything about problems with the sample batches. You need to tell me that you're having problems when you're collecting the sample, not when it's done. This firm did not have anything documented…. If you don't write anything down, it's very hard to do a failure investigation."

A particular problem that occurs during some inspections is when analysts go back in the records and change sample weights. Rolli said that "sample weights have to be documented. It's unacceptable for an analyst to put in a weight later. It's unacceptable to go back in and change it later on."

Another problem noted by Rolli is when analysts do not process rejected sample batches. "You want to make sure that you always have a result. We are looking at how the data is going in and how is it going out."

- Joanne S. Eglovitch ([email protected])