4.5 Laboratory Quality Assurance, Chain of Custody, and Desorption Efficiency

Laboratory Quality Assurance, Chain of Custody, and Desorption Efficiency

An exposure result is a legal and medical document. It may determine whether an employer is cited, whether a worker is removed from a job, or whether a plaintiff prevails twenty years later. The quality system described in this section is what separates a defensible number from a number that a competent opposing expert will destroy.

---

1. Desorption Efficiency

Sorbent sampling has an inherent loss: not all of the analyte adsorbed onto charcoal or silica gel comes back off in the desorbing solvent. Desorption efficiency (DE) quantifies that loss.

Determination

A known mass of analyte is spiked directly onto the same lot of sampling medium, allowed to equilibrate (typically overnight), then desorbed and analysed exactly as a field sample would be:

$\text{DE} = \frac{\text{mass recovered from the spiked medium}}{\text{mass spiked}}$

DE is determined at several loading levels bracketing the expected field mass, because it is not constant across the working range — it typically falls at very low loadings where irreversible adsorption is proportionally more significant.

Correction

Every reported field mass must be corrected:

$\text{Corrected mass} = \frac{\text{mass found by analysis}}{\text{DE}}$

Failing to apply the correction biases every result low, which is the direction that matters: it under-reports exposure.

Worked example. A charcoal tube is analysed and 0.180 mg of toluene is found on the front section. The laboratory reports a desorption efficiency of 0.90 for toluene on that lot. The corrected mass is 0.180 / 0.90 = 0.200 mg. If the sample volume was 12 L, the concentration is 0.200 mg / 0.012 m³ = 16.7 mg/m³ — an 11% difference from the uncorrected value.

Acceptance

A DE around 0.90 to 1.00 is typical for a well-matched analyte/sorbent/solvent combination. NIOSH methods generally expect DE at or above about 0.75. Below that, the correction factor is doing too much work: small variations in DE propagate into large errors in the reported result, and the correct response is to reconsider the medium or the desorbing solvent, not to apply a factor of 2.

---

2. Blanks: Three Kinds, Three Purposes

Field blanks are the workhorse. The usual rule is at least 10% of the sample set, with a minimum of two and typically no more than ten per sample set. The mean field blank mass is subtracted from each sample mass before the DE correction is applied. A field blank carrying substantial analyte invalidates the set — you cannot subtract your way out of a contaminated sampling process.

Spiked (control) samples and duplicates complete the picture: spikes verify accuracy against a known quantity, and side-by-side duplicates verify precision.

---

3. Chain of Custody

Chain of custody is the documented, unbroken record of who possessed each sample from the moment of collection until analysis and disposal. A sample is "in custody" when it is in your physical possession, in your view, secured by you, or in a designated secure area.

A defensible chain of custody record contains:

1. Unique sample identifier and collection date, time, and location
2. Sampler's name and signature
3. Analyte requested and analytical method specified
4. Sample matrix, medium type, and lot number
5. Volume sampled, or flow rate and duration
6. Preservation applied and required holding time
7. Every transfer, with relinquishing and receiving signatures, date, and time
8. Condition of the shipping container and seals on receipt at the laboratory

+-------------------------------------------------------------+
|  COLLECTION -> SEALED -> SHIPPED -> RECEIVED -> ANALYSED     |
|      sign        seal      sign       sign        sign       |
|                                                              |
|  A gap anywhere in this line is a gap in the defensibility   |
|  of every number that comes out of the laboratory.           |
+-------------------------------------------------------------+

Custody seals on the shipping container let the receiving laboratory certify that the samples were not accessible in transit. A broken seal is documented as a discrepancy, not quietly ignored.

---

4. Preservation and Holding Times

Sample integrity degrades with time and temperature. Common requirements:

• Refrigeration for thermally labile analytes and for most bioaerosol and derivatised samples
• Protection from light for photodegradable analytes
• Prompt analysis of hexavalent chromium and derivatised isocyanate samples
• Sealed, capped, and separately bagged sorbent tubes to prevent cross-contamination from high-concentration samples in the same shipment

Exceeding a method holding time is a data qualification, not a discretionary matter: the result is flagged, and its usability for a compliance or medical decision must be assessed explicitly.

---

5. Laboratory Competence

Two independent lines of evidence establish that a laboratory can do the work:

• Accreditation. The AIHA Laboratory Accreditation Programs (AIHA-LAP) accredit laboratories against ISO/IEC 17025 for defined fields of testing — industrial hygiene, environmental lead, environmental microbiology. Accreditation is scope-specific: a laboratory accredited for lead is not thereby accredited for asbestos.
• Proficiency testing. Blind samples of known concentration are analysed at intervals and scored against reference values. The Industrial Hygiene Proficiency Analytical Testing (IHPAT) program and its analogues for asbestos and lead provide the ongoing performance record.

Some regulatory standards require accreditation and proficiency participation explicitly, and for enforcement or litigation purposes the absence of either is a straightforward line of attack on the data.

---

6. Limits of Detection and Quantitation

Two figures of merit bound what the number means:

• Limit of detection (LOD): the smallest mass distinguishable from the blank, conventionally three times the standard deviation of the blank response.
• Limit of quantitation (LOQ): the smallest mass that can be reported with acceptable precision, conventionally ten times that standard deviation.

A result reported as "less than the LOD" is not zero. For statistical treatment of censored data, a common convention is to substitute LOD divided by the square root of 2, though maximum-likelihood and regression-on-order-statistics methods are preferred when a substantial fraction of results are non-detects. Reporting non-detects as zero biases the mean low and understates exposure.