延续注册要准备什么：有效期到期前的时间倒排

结论：延续注册的难点不在难度，在时间窗口

延续注册本身不复杂，但它有一个刚性约束：申请要在注册证有效期届满前的规定期限内提出，错过窗口就只能按首次注册重新走。 两者的工作量差着数量级。

实际出问题的项目，很少是因为技术上做不到，多数是因为启动太晚。等想起来要延续时，发现有几项数据需要补测，而检测周期加上资料整理已经压不进剩余的时间。

从到期日往前倒排

下面这个时间轴是按常见情形给的参考量级，具体要按产品复杂度和实验室档期调整。

看起来时间很宽裕，但第一行和第二行之间往往会卡住。梳理变更需要研发、生产、质量几个部门配合，而这件事没有明确的责任人时，很容易拖两三个月。建议把它当成一个有交付物的任务来管理，交付物就是一份变更清单和一份标准符合性核对表。

哪些情形需要补充检测

延续注册不必然重新送检，但以下几种情形通常要补：

执行标准换版。 有效期内产品执行的标准发生了换版，而新版本的要求或方法有实质变化。这时需要按新版本补测受影响的项目。哪些项目受影响，要逐条比对新旧版本的差异，不能笼统地全测或全不测。

有效期内发生过变更但未单独申报。 有些企业把小变更累积起来，打算在延续时一并处理。这种做法的风险是，累积的变更合在一起可能已经超出「小变更」的范围，需要按变更注册的口径补齐证据。

原报告的方法已被废止。 老产品的原始报告可能执行的是早已废止的方法。方法废止不等于报告作废，但延续时往往需要用现行方法补充确认。

监管关注点发生变化。 某一类产品如果近年出现过共性问题，相关项目可能被重点关注，即使标准没变也可能需要补充数据。

标准符合性核对怎么做

这一步是延续准备里最容易做得敷衍的环节。建议做成一张逐条对照表：

把原注册时执行的标准清单列出来，逐个查现行有效版本；有换版的，比对新旧差异；差异涉及技术要求或试验方法的，标记为需要评估；评估后确定补测项目。

核对标准现行版本这件事，不要依赖记忆或旧资料。 标准的换版信息应当从正式渠道确认，站内的标准页面、供应商提供的资料、几年前的培训材料都不构成依据。这一条我们自己也踩过坑，所以写在这里。

有效期内发生过变更怎么处理

如果产品在有效期内做过变更，处理方式取决于变更当时有没有单独申报。

已单独申报并获批的，延续时按变更后的状态准备资料，原始报告与变更报告一并提供。

未单独申报的，要先判断这些变更是否本应单独申报。属于应当申报而未申报的，需要在延续时补齐相应证据，并做好被追问的准备。确实属于无需申报范畴的，在延续资料中说明即可。

把变更历史梳理清楚，比补几项检测更重要。 检测可以花钱买周期，变更历史的梳理只能靠内部记录，记录不全的时候补不回来。

常被遗漏的几类准备工作

除了检测数据，延续注册还有几项准备容易被漏掉：

生产工艺的一致性说明。 产品在有效期内如果工艺有调整，即使产品本身没变，也可能需要说明工艺调整对产品特性的影响。

说明书与标签的现行版本。 这几年说明书和标签的要求变化较多，沿用几年前的版本提交，容易在形式审查阶段被退。

质量管理体系的运行证据。 延续时通常需要体现体系持续有效运行，包括内审、管理评审、不良事件监测的记录。

不良事件与投诉的汇总分析。 有效期内收到的投诉和不良事件要汇总，并说明处理情况。如果其中涉及性能问题，可能需要对应的检测支撑。

这几项的共同点是：都不是临到期才能做的事，但都需要平时有积累。 平时不记录，到期时补不出来。

时间压缩时的取舍

如果确实启动晚了，时间不够，需要做取舍。优先级建议这样排：

先保证法规要求必须提交的项目；其次是标准换版涉及的实质性差异项；再次是变更相关的补充数据；最后才是锦上添花的补充材料。

检测这一环如果压不进去，可以考虑与实验室沟通加急，但要提前了解加急的可行性——有些项目的周期是物理决定的，比如老化、耐久，加钱也压缩不了。 这类项目一旦漏排，就只能接受延后申报的后果。

一个实务提醒

延续申请提交之后到批准之间还有一段时间，这段时间产品能否继续生产销售，要按当时的规定执行。建议在启动延续工作时就把这一点确认清楚，并据此安排生产和库存计划。有企业按「提交即可继续」的理解安排了生产，后来发现口径不同，造成了库存积压。这类信息应当从正式渠道确认，不要依赖同行经验。

我们的做法

我们接延续类委托时，会先看两份材料：原注册时的检测报告，以及有效期内的变更清单。看原报告是为了确认执行的标准版本和方法；看变更清单是为了判断有没有牵动到检测结论。

比较常见的情形是，委托方以为只要按新版标准全测一遍就稳妥，我们比对下来发现新旧版本的实质差异只涉及少数几项，其余可以沿用。逐条比对省下来的项目，通常比想象的多。 反过来也有：委托方觉得没变化，比对后发现某项方法已经改了判定依据，原数据不再适用。

如果你有产品临近到期但不确定要补哪些项目，可以把原报告和变更清单发过来一起核，或者直接联系：132 4819 8029。能力范围见服务介绍，常见问题见常见问题，更多内容见知识库。

English version

Conclusion: renewal is not technically hard, but the window is rigid

Renewal of a device registration is not complex work. What makes it difficult is a fixed constraint: the application must be filed within a defined period before the certificate expires. Miss that window and the product goes back through initial registration, which is an order of magnitude more work.

Projects that run into trouble rarely do so because something was technically impossible. They run into trouble because they started late. By the time renewal comes to mind, a few data gaps have been identified, and the testing turnaround plus document preparation no longer fits in the time remaining.

Working backwards from the expiry date

As a rough guide: start the assessment about twelve months out by compiling every change made during the certificate's validity and checking the current versions of the applicable standards. By around ten months out, decide whether supplementary testing is needed and fix the item list. Submit samples at about nine months. Have reports in hand by six months and begin assembling the dossier. Complete internal review by four months. File within the prescribed period.

This looks generous, but the gap between the first and second steps is where projects stall. Compiling the change history requires cooperation from R&D, production and quality, and without a named owner it commonly drifts for two or three months. It helps to treat it as a task with a deliverable: a change list and a standards conformity checklist.

When supplementary testing is needed

Renewal does not automatically require retesting, but several situations usually do.

A standard has been revised during the validity period and the new version changes requirements or methods substantively. Affected items are retested. Which items are affected should be determined by a clause-by-clause comparison, not by retesting everything or nothing.

Changes were made during the validity period but not separately filed. Some manufacturers accumulate minor changes intending to handle them at renewal. The risk is that the accumulated changes together exceed what counts as minor.

The method used in the original report has been withdrawn. Withdrawal does not invalidate the existing report, but confirmation under a current method is often needed at renewal.

Regulatory attention has shifted. Where a product category has shown common problems, related items may attract scrutiny even if the standard itself has not changed.

Checking standards conformity

This step is the one most often done superficially. Build a clause-by-clause table: list the standards applied at original registration, check the current valid version of each, compare old and new where a revision has occurred, flag any difference touching requirements or methods, and decide on retesting from there.

Do not rely on memory or old documents to establish which version is current. Revision information should be confirmed through formal channels. Website pages, supplier-provided summaries and training material from a few years ago are not evidence. We have been caught by this ourselves, which is why it is stated here.

Handling changes made during the validity period

If changes were filed and approved separately, prepare the renewal on the post-change configuration and provide both the original and the change-related reports.

If they were not filed, first judge whether they should have been. Where filing was required and did not happen, the supporting evidence needs to be assembled at renewal, and questions should be expected. Where filing genuinely was not required, a statement in the renewal dossier is usually sufficient.

Getting the change history straight matters more than adding a few test items. Testing can be bought with money and schedule; change history can only come from internal records, and when those are incomplete they cannot be reconstructed.

How we handle it

For renewal work we ask for the original test report and the change list. The report tells us which standard versions and methods were applied; the change list tells us whether anything has disturbed the original conclusions.

A frequent outcome is that the client assumed a complete retest against the new standard version was the safe option, and our comparison shows the substantive differences affect only a few items while the rest can stand. The savings from a clause-by-clause comparison are usually larger than expected. The reverse also occurs: a client sees no change, and comparison reveals that a method's acceptance basis has been revised so the original data no longer applies.

If you have a product approaching expiry and are unsure what needs to be retested, send us the original report and the change list. Phone or WeChat: +86 132 4819 8029.