The medical device industry is in the process of shifting from the adoption of ISO 594 to ISO 80369 as the preferred standard for qualifying luer connectors, as FDA will no longer accept ISO 594 in support of premarket submissions after December 31, 2019. Image courtesy of DDL Inc.
In the years since ISO 80369 was announced, the impact of the changes to the medical device industry have begun to sink in, and momentum has been steadily building as designers are adopting the changes. At DDL, we have observed a dramatic shift away from ISO 594 as the preferred standard for qualifying luer connectors, which is inevitable, but surprising in terms of the pace of adoption within the industry as a whole.
As with every new standard, there are certain changes that have more impact than the rest and generate the bulk of the questions that come up during new project discussions. Below are responses to the most common questions we receive from our customers. As implementation of the standard progresses throughout the industry, I would like to continue to update the audience by answering new questions as they arise in a future article.

 
What are the requirements for dimensional compliance?
Dimensional conformance to the figures and tables in Annex B of ISO 80369 (Part 3 for enteral connectors, Part 5 for limb cuff connectors, Part 6 for neuraxial connectors, and Part 7 for IV and hypodermic connectors) is a requirement. In your submission, you will be required to show objective evidence that your molded or machined connector meets the dimensional specifications for the connector style. If the connector does not meet the specs, then it is prohibited from being labeled as a connector per the intended application category (e.g., a luer connector).
Many different measurement techniques may be employed to fully and accurately assess a connector’s dimensions, including contact and vision CMM, CT scanning, potting and sectioning, calculation by interpolation, and using reproduction rubber to create casts of internal features. The specific techniques depend on the connector type and gender and the judgment of the inspection lab.
As manufacturing processes pass the design verification hurdle and get into commercial production, a different concern has started to arise – that of in-process checks. Commonly, the go/no-go gauging test using the plug or ring gauge per Figure 3A, 3B, or 3C in ISO 594-1 is used as a simple and effective means to monitor for process drift. However, as this requirement doesn’t exist in ISO 80369-7, the suitability of this method has been brought into question. The surprising fact is that the tolerances for the luer taper are actually looserin ISO 80369-7 than in ISO 594-1, so utilizing the go/no-go gauge provides an additional layer of assurance that the samples coming off the mold are within specification.
Is non-interconnectability testing a requirement?
There is an update pending to ISO 80369-1 that makes this point clear – if a connector meets the material (modulus), dimensional, and performance requirements of one of the existing application categories, then it is considered to be non-interconnectable by default. The non-interconnectability characteristics of the connector designs were validated by the ISO committee responsible for publishing the standards; so as long as the critical assumptions are adhered to, it is not required to perform the non-interconnectability test per ISO 80369-1 Annex B.
For non-standard geometries, there is a special case regarding luer access devices, or LADs. These are female lock connectors that are designed to mate to male luers but that have an internal valve, rather than a luer taper that is actuated by the male taper and opens the fluid path. LADs do not meet the definitions for a luer and are therefore outside the scope of ISO 80369-7. However, demonstrating conformance with the performance requirements and certain dimensional elements that support non-interconnectability characteristics are meaningful and help support claims of compatibility with male luer connectors.
Are variable test methods worth the effort?
Like many complex questions, the answer to this one is, “it depends.” The performance tests described in Annexes B through I in ISO 80369-20 are all set up as attribute analyses. Annex J in the standard describes ways in which the test procedures may be modified in order to perform them so as to permit variable analysis. The advantage to variable methods is that there is typically more power to each data point, meaning fewer samples need to be tested in order to achieve the required confidence and reliability intervals. Each method requires a different degree of modification, ranging from trivial to fairly complex, in order to make the conversion. In loose terms, here’s how I would rank the relative degree of difficulty for each test method:
Trivial
简单
• Leakage by pressure decay.
压差泄漏
• Sub-atmospheric pressure air leakage.
负压气体泄漏
• Disconnection by unscrewing.
脱模断开
Moderate
中等
• Resistance to separation from unscrewing.
卸下时的分离阻力
Challenging
困难
• Positive pressure liquid leakage.
正压液体泄漏
• Resistance to separation from axial load.
受轴向载荷的分离阻力
• Resistance to overriding.
重载阻力
The “Stress Cracking” method can’t be modified in order to obtain variable data, which is unfortunate, because that is the most time intensive test of the lot.
For the methods that I place in the “Trivial” category, the only modification necessary is in the data analysis, because the standard test output is a numerical value. The first step is to test for normality and transform the data to normal, if necessary. From there, the upper or lower specification limit can be determined on the sample group and compared to the limit prescribed in the standard for the connector application being tested.
For the “Resistance to Separation from Unscrewing” test, the main challenge is in ensuring that the equipment used to measure the unscrewing torque has the necessary control required for the test. For example, spring-loaded torque watches are not suitable because the torque reading can get distorted when the sample suddenly breaks loose from the reference connector. The reading is also sensitive to the torque application rate, so applying some measures to ensure consistency are important for repeatability and reproducibility.
The tests in the “Challenging” category require equipment that is substantially different than what is commonly used for the attribute methods. The “Resistance to Separation from Axial Load” and “Resistance to Overriding” tests commonly require special fixturing in order to securely grip the sample in tensile or rotational test systems, adding time and expense to the project for design and qualification work. Also of great concern is the prospect of damage to the reference connectors. A sufficiently robust sample connector design could require application of forces that cause localized yielding of the stainless steel.
All in all, variable methods can be a great option in situations where high risk attributes mandate very high confidence and reliability, or when the sample connectors are expensive or in limited supply, but there are also significant complications that need to be addressed. In many cases, the value of performing variable analysis is more difficult to realize than it initially appears.