Hidden Cracks:
The Four-Year
PRPE Integrity Study
A landmark longitudinal investigation reveals that nearly half of all radiation protective garments develop hidden defects — and that new, repaired, and everyday-use items are equally at risk.
The Assumption That Lead Aprons Work
In every fluoroscopy suite, cath lab, and surgical theatre across the world, staff don lead aprons before procedures — and proceed on the assumption that those garments are doing their job. This assumption is foundational to occupational radiation protection. Without it, every millimetre of stated lead equivalence, every inspection sticker, every procurement decision becomes meaningless.
Yet the integrity of personal radiation protective equipment (PRPE) — meaning the physical condition of the attenuating layer inside the garment — has received remarkably little rigorous scientific attention. Prior work had described snapshot data from single-year inspections, or qualitative observations without systematic tear quantification. No study before Kellens et al. had followed a large cohort of PRPE longitudinally, tracking the progression of cracks and tears year on year across a real clinical environment.
This study from Ghent University and AZ Sint-Jan hospital in Belgium fills that gap. Over four consecutive years from 2018 to 2021, every piece of PRPE belonging to a large general hospital was inspected annually using fluoroscopic X-ray imaging — the only reliable method for detecting defects invisible to the naked eye.
How Do You Find a Hidden Tear?
Each piece of PRPE — aprons, vests, skirts, and thyroid shields — was examined fluoroscopically on a Siemens Luminos dRF table using a standardised abdomen AP protocol. Defects appear as bright white areas in fluoroscopy mode, indicating gaps in the attenuating layer. An X-ray-opaque grid with 1 cm interspace was used to measure tear areas, and a tennis ball was placed beneath suspicious areas to improve visualisation.
Rejection followed the Lambert & McKeon criteria — total defect area exceeding 670 mm² — supplemented by an institutional rule that any single tear above 15 mm² warranted immediate rejection. Thyroid shields were held to a stricter standard (11 mm² total, 5 mm² individual). Every piece was assigned a unique tracking number to enable true longitudinal follow-up.
Over four years, the team conducted 2,588 quality control checks on 1,011 unique pieces — making this the largest and most rigorous longitudinal PRPE integrity dataset published to date.
The Numbers That Should Change Practice
Across the 4-year study period, 55% of all quality control checks revealed tears. Of those, 34% exceeded rejection criteria — giving an average annual rejection rate of 18.6%. In terms of individual pieces, 47.3% of all PRPE showed at least one tear at some point, with 31% of those needing to be condemned.
| PRPE Type | Rejection Rate | Median Tear Area | Notable |
|---|---|---|---|
| Thyroid shields | 8.1% | 14 ± 34 mm² | Smallest — but tears can exceed 160 mm² |
| Skirts | 26.3% | 40 ± 117 mm² | Highest rejection — sitting folds attenuating layer |
| Aprons | 19.6% | 23 ± 64 mm² | No significant difference from vests |
| Vests | 24.0% | 30 ± 69 mm² | Second highest rejection rate |
Skirts showed the highest rejection rate and largest tear areas — a finding the authors attribute to the repeated folding of the attenuating layer that occurs when staff sit while wearing them. Improper storage, where garments are folded rather than hung, produces the same crease-and-weaken pattern.
"New garments are not guaranteed to remain free of cracks in their first year of use — greatly emphasising the requirement of PRPE quality assurance starting from day one."
Kellens et al. — Insights into Imaging, 2022New Doesn't Mean Safe
One of the study's most clinically important findings concerns newly purchased equipment. Of 287 newly registered pieces of PRPE from different manufacturers, 6% showed tears within their very first year of service. Of those newly defective items, 88.2% exceeded rejection criteria and required immediate condemnation.
The median tear area in newly defective pieces was 70 ± 60 mm² — well above the 15 mm² single-tear threshold. This finding invalidates a common assumption in procurement practice: that a newly purchased garment can be placed into service without any baseline quality check or early follow-up inspection.
The authors note similar findings in Glaze et al. (1984) and Oppliger-Schäfer et al. (2009), both of whom documented defects in new items. The evidence is not new — but it has now been quantified longitudinally at scale for the first time.
Repair Is Not a Solution
Of 766 quality control checks performed on repairable pieces, 31.7% led to at least one repair. The study then tracked what happened next: nearly 50% of repaired pieces were rejected again in the following year's inspection. In specific two-year windows, rejection-after-repair rates reached 60%.
The explanation is mechanistic. Figure 2 of the paper tracks a single vest through three consecutive repair cycles over four years. Each repair creates a patch — and tears reliably develop directly adjacent to those patches, apparently driven by material stress concentrations at the boundary between original and patched material. By 2020, after two repairs, the vest showed a tear of approximately 300 mm². After three repairs in 2021, another large tear had formed beside the newest patch.
Repair appears to shift rather than resolve the structural weakness. Departments that send rejected garments for repair and return them to service with confidence are, statistically, taking a 50/50 gamble.
Brand Is the Strongest Predictor of Rejection
Rejection rates varied dramatically by manufacturer. The repairable brands (A and B in the study — Infinity and Tema) showed average rejection rates of 27% and 21.9% respectively. In contrast, Brand E (Protec X) showed an average rejection rate of just 1.1% across the study period.
The pain clinic and cardiology departments showed the highest rejection rates — consistent with the intensive, close-proximity fluoroscopy work these roles demand. The correlation between usage intensity and defect rate is clear, though the authors caution that brand confounding (certain brands may be concentrated in particular departments) complicates direct comparison.
What Needs to Change
- Annual X-ray-based inspection of all PRPE is the minimum standard — visual inspection alone is insufficient to detect attenuating layer defects
- New garments must be inspected on receipt and within their first year of service — manufacturer quality is not guaranteed
- Repaired PRPE should be re-inspected within 12 months and should not be considered a reliable long-term solution
- Skirts deserve particular attention given their higher rejection rate — storage on hangers (not folded) should be standard
- Brand selection is a legitimate quality consideration: rejection rates vary by up to 26 percentage points between manufacturers
- Future rejection criteria should be updated to reflect dosimetric evidence rather than the economically-derived thresholds from Lambert & McKeon (2001)
Hidden Cracks:
The Four-Year
PRPE Integrity Study
A landmark longitudinal investigation reveals that nearly half of all radiation protective garments develop hidden defects — and that new, repaired, and everyday-use items are equally at risk.
The Assumption That Lead Aprons Work
In every fluoroscopy suite, cath lab, and surgical theatre across the world, staff don lead aprons before procedures — and proceed on the assumption that those garments are doing their job. This assumption is foundational to occupational radiation protection. Without it, every millimetre of stated lead equivalence, every inspection sticker, every procurement decision becomes meaningless.
Yet the integrity of personal radiation protective equipment (PRPE) — meaning the physical condition of the attenuating layer inside the garment — has received remarkably little rigorous scientific attention. Prior work had described snapshot data from single-year inspections, or qualitative observations without systematic tear quantification. No study before Kellens et al. had followed a large cohort of PRPE longitudinally, tracking the progression of cracks and tears year on year across a real clinical environment.
This study from Ghent University and AZ Sint-Jan hospital in Belgium fills that gap. Over four consecutive years from 2018 to 2021, every piece of PRPE belonging to a large general hospital was inspected annually using fluoroscopic X-ray imaging — the only reliable method for detecting defects invisible to the naked eye.
How Do You Find a Hidden Tear?
Each piece of PRPE — aprons, vests, skirts, and thyroid shields — was examined fluoroscopically on a Siemens Luminos dRF table using a standardised abdomen AP protocol. Defects appear as bright white areas in fluoroscopy mode, indicating gaps in the attenuating layer. An X-ray-opaque grid with 1 cm interspace was used to measure tear areas, and a tennis ball was placed beneath suspicious areas to improve visualisation.
Rejection followed the Lambert & McKeon criteria — total defect area exceeding 670 mm² — supplemented by an institutional rule that any single tear above 15 mm² warranted immediate rejection. Thyroid shields were held to a stricter standard (11 mm² total, 5 mm² individual). Every piece was assigned a unique tracking number to enable true longitudinal follow-up.
Over four years, the team conducted 2,588 quality control checks on 1,011 unique pieces — making this the largest and most rigorous longitudinal PRPE integrity dataset published to date.
The Numbers That Should Change Practice
Across the 4-year study period, 55% of all quality control checks revealed tears. Of those, 34% exceeded rejection criteria — giving an average annual rejection rate of 18.6%. In terms of individual pieces, 47.3% of all PRPE showed at least one tear at some point, with 31% of those needing to be condemned.
| PRPE Type | Rejection Rate | Median Tear Area | Notable |
|---|---|---|---|
| Thyroid shields | 8.1% | 14 ± 34 mm² | Smallest — but tears can exceed 160 mm² |
| Skirts | 26.3% | 40 ± 117 mm² | Highest rejection — sitting folds attenuating layer |
| Aprons | 19.6% | 23 ± 64 mm² | No significant difference from vests |
| Vests | 24.0% | 30 ± 69 mm² | Second highest rejection rate |
Skirts showed the highest rejection rate and largest tear areas — a finding the authors attribute to the repeated folding of the attenuating layer that occurs when staff sit while wearing them. Improper storage, where garments are folded rather than hung, produces the same crease-and-weaken pattern.
"New garments are not guaranteed to remain free of cracks in their first year of use — greatly emphasising the requirement of PRPE quality assurance starting from day one."
Kellens et al. — Insights into Imaging, 2022New Doesn't Mean Safe
One of the study's most clinically important findings concerns newly purchased equipment. Of 287 newly registered pieces of PRPE from different manufacturers, 6% showed tears within their very first year of service. Of those newly defective items, 88.2% exceeded rejection criteria and required immediate condemnation.
The median tear area in newly defective pieces was 70 ± 60 mm² — well above the 15 mm² single-tear threshold. This finding invalidates a common assumption in procurement practice: that a newly purchased garment can be placed into service without any baseline quality check or early follow-up inspection.
The authors note similar findings in Glaze et al. (1984) and Oppliger-Schäfer et al. (2009), both of whom documented defects in new items. The evidence is not new — but it has now been quantified longitudinally at scale for the first time.
Repair Is Not a Solution
Of 766 quality control checks performed on repairable pieces, 31.7% led to at least one repair. The study then tracked what happened next: nearly 50% of repaired pieces were rejected again in the following year's inspection. In specific two-year windows, rejection-after-repair rates reached 60%.
The explanation is mechanistic. Figure 2 of the paper tracks a single vest through three consecutive repair cycles over four years. Each repair creates a patch — and tears reliably develop directly adjacent to those patches, apparently driven by material stress concentrations at the boundary between original and patched material. By 2020, after two repairs, the vest showed a tear of approximately 300 mm². After three repairs in 2021, another large tear had formed beside the newest patch.
Repair appears to shift rather than resolve the structural weakness. Departments that send rejected garments for repair and return them to service with confidence are, statistically, taking a 50/50 gamble.
Brand Is the Strongest Predictor of Rejection
Rejection rates varied dramatically by manufacturer. The repairable brands (A and B in the study — Infinity and Tema) showed average rejection rates of 27% and 21.9% respectively. In contrast, Brand E (Protec X) showed an average rejection rate of just 1.1% across the study period.
The pain clinic and cardiology departments showed the highest rejection rates — consistent with the intensive, close-proximity fluoroscopy work these roles demand. The correlation between usage intensity and defect rate is clear, though the authors caution that brand confounding (certain brands may be concentrated in particular departments) complicates direct comparison.
What Needs to Change
- Annual X-ray-based inspection of all PRPE is the minimum standard — visual inspection alone is insufficient to detect attenuating layer defects
- New garments must be inspected on receipt and within their first year of service — manufacturer quality is not guaranteed
- Repaired PRPE should be re-inspected within 12 months and should not be considered a reliable long-term solution
- Skirts deserve particular attention given their higher rejection rate — storage on hangers (not folded) should be standard
- Brand selection is a legitimate quality consideration: rejection rates vary by up to 26 percentage points between manufacturers
- Future rejection criteria should be updated to reflect dosimetric evidence rather than the economically-derived thresholds from Lambert & McKeon (2001)
Start writing here...
