OOS-附件2A.doc

201410 PA/PH/OMCL (14) 89 结果的评估和报告---附录2A 定量检测项目的复验程序举例
OMCL Network of the Council of Europe
QUALITY ASSURANCE DOCUMENT
PA/PH/OMCL (14) 89
Evaluation and Reporting of Results
结果的评估和报告
Annex 2A
Full document title and reference
Evaluation and reporting of Results---Annex 2A Examples of Re-test Programmes for Quantitative Tests PA/PH/OMCL (14) 89
全名和索引
结果的评估和报告---附录2A 定量检测项目的复验程序举例PA/PH/OMCL(14)89
Document type
Guideline
文件类型
指南
Legislative basis
/
法规依据
无
Date of first adopt
August 2014
首次采纳日期
2014年8月
Date of original entry into force
October 2014
首次实施日期
2014年10月
Date of entry into force of revised document
/
修订版生效日期
无
Previous titles/other references/last valid version
/
之前题目/其它索引号/上一有效版本
无
Custodian Organisation
The present document was elaborated by the OMCL Network/EDQM of the Council of Europe
责任机构
现行文件由欧盟委员会EDQM/OMCL网络编制
Concerned Network
GEON
相关网络
GEON

ANNEX II A OF THE OMCL NETWORK GUIDELINE
“EVALUATION AND REPORTING OF RESULTS”
EXAMPLES OF RE-TEST PROGRAMMES FOR QUANTITATIVE TESTS
TABLE OF CONTENT
INTRODUCTION 概述
Approach 1: Active pharmaceutical ingredient, 2 initial determinations
方法1：原料药，2次初始检测
Approach 2: Active pharmaceutical ingredient, 3 initial determinations
方法2：原料药，3次原始检测
Approach 3: Impurity tests (e.g. Related substances by HPLC)
方法3：杂质检测 （例如HPLC检测有关物质）
Approach 4: Finished products
方法4：制剂检测
Approach 5: Products with insufficient validation data
方法5：验证数据不充分的产品
Approach 6: Re-test programme based on statistical assay layouts (Ph. Eur. 5.3)
方法6：根据统计学含量布局制订的复验计划（欧洲药典5.3）
Approach 7: Re-test programme based on known intermediate precision
方法7：根据已知中间精密度的复验程序
Approach 8: Approach for cases of unexplained lack of repeatability
方法8：缺乏重复性又不能解释情况下所用的方法

INTRODUCTION 概述
This document is an Annex to the core document “Evaluation and reporting of results”, PA/PH/OMCL (13) 113(in its current version), and it should be used in combination with it when planning, performing and documenting the evaluation process and reporting of results of quantitative tests with the expected Gaussian distribution.
本文件是“结果的评估与报告”核心文件PA/PH/OMCL (13) 113（现行版本）的附录。在计划、实施和记录符合正态分布的定量测试结果的评估过程和报告结果时要与核心文件一起使用。
The core document contains the Introduction, Scope and General requirements for the evaluation of results (in routine cases or otherwise) and the reporting of results.
核心文件包括结果评估（在常规情形还是其它情形和报告的概述、范围和通用要求。
The proposed approaches are only relevant if the decision is based on observed repeatability[1]. In all cases, the laboratory should make a decision based on documented and sound scientific judgement. This Annex should not be taken as a list of compulsory requirements. It is left to the professional judgement and background experience of each OMCL to decide on the most relevant procedures to be undertaken in order to give evidence that the evaluation and reporting of results is well managed. Although two or three initial determinations are described in most examples of this Annex, each OMCL is responsible for deciding what is the minimum number of replicates that should be taken into account for a certain test.
如果基于观察到的重复性来做出决定，则提议的方式仅是相关而已。在所有情形下，化验室应根据书面的科学合理判断来做出决定。本附录不应作为是强制要求的清单，每个OMCL应该根据专业判断和背景知识来决定采用最相关的程序，以提供证据证明对结果的评估和报告有良好的管理。虽然在本附录的大多数例子中描述了2次或3次初始检测，但每个OMCL应决定在某个具体的测试中需考虑的平行检测的最少次数。
This Annex presents several examples of the evaluation of results for quantitative testing of medicinal products, which could be set up in combination with the general requirements given in the core document. The examples are not intended to be all-inclusive, and other valid approaches may be adopted for evaluation of the acceptability of test results.
本附录给出了几个药品定量检测中结果评估的例子，它们可以与核心文件中的通用要求结合来看。这些例子无意列举出所有可能，在检测结果的可接受度评估中也可以采用其它有效方法。
This document is based on publications in Pharmeuropa Vol. 9, No. 1, 148-156 (1997) and Pharmeuropa Vol. 11, No. 4, 571-577 (1999). The proposals for approach 1, 2 and 4 were tested against datasets obtained from proficiency tests and have been shown to be satisfactory for making a decision.
本文件是依据欧洲药典在线卷9第1篇页148-156（1997）和欧洲药典在线卷11第4篇页571-577（1999）而制订的。方法1、2和4所用的方案经过了专业测试所获得的数据系列的测试，显示很令人满意，可以用以做出决策。
The following table gives an overview about the described possible approaches and the situations where they may be applied, see below:
下表给出了所述可能方法的概览，以及这些方法可能适用的情形。
Approach
方法
Title
标题
Situation
情形
1
Active pharmaceutical ingredient, 2 initial determinations
This approach is applicable for APIs only

原料药，2次初始检测
本方法仅适用于原料药
2
Active pharmaceutical ingredient, 3 initial determinations
This approach is applicable for APIs only

原料药，3次初始检测
本方法仅适用于原料药
3
Impurity tests (e.g.: Related substances by HPLC)
This approach is intended for trace level tests and may be used for quantitative impurity tests, where an analytical/instrumental response is obtained (peak area, for example)

杂质检测（例如，HPLC检测有关物质）
本方法用于痕量检测，可以用于利用分析方法/仪器响应（例如峰面积）来进行定量杂质检测的情况
4
Finished products
This approach is intended for finished products with fully validated methods regarding repeatability and intermediate precision described in the MA file. It is not intended for analyte concentrations at trace level.

制剂
本方法适用于经过完整验证的制剂检测方法，且方法的重复性和中间精密度在MA文件中有载。不适用于痕量水平的分析物浓度检测。
5
Products with insufficient validation data
This approach is to be regarded as a tool to establish an acceptance criterion for precision of the replicates of a certain test, when the laboratory has no other mean to evaluate the dispersion of the results, as the available validation data are scarce or there are no validation data at all. It is not intended for analyte concentrations at trace level.

验证数据不充分的产品
本方法是作为一种工具，当化验室没有其它方法来评估结果的分散度，且已有验证数据很少，或完全没有验证数据时，用来建立一个方法重复检测时的精密度的可接受标准。
6
Retest programme based on statistical assay layouts (Ph. Eur. 5.3)
Results from bioassays in statistical layout according to Ph. Eur. 5.3. For combination of assays it is desirable that the individual results are obtained in identical or similar assay layouts.

基于统计学含量布局的复验计划（欧洲药典5.3）
根据药典5.3统计学生效含量所得的结果。对于含量结果的合并，最好单个结果具有相同或类似的含量布局
7
Retest programme based on known intermediate precision
This approach is intended for assays with well-known in-house intermediate precision, e.g. in OCABR routine analysis

根据已知中间精密度的复验计划
本方法适用于已知内部中间精密度的含量，例如，在OCABR常规分析中
8
Approach for cases of unexplained lack of repeatability
For products where sample recovery problems are encountered or there is no plausible reason for lack of precision

缺乏重复性且不能解释的情形所用方法
用于样品有回收问题的药品，或缺乏精密度没有貌似有理的原因

Approach 1: Active pharmaceutical ingredient, 2 initial determinations
方法1：原料药，2次初始检测
Perform two determinations. If the RSD2[2] is smaller than the RSDmax permitted for two determinations (seeTable 1), and the mean falls within the content limits, the sample passes. If either of these two conditions is not met, one further determination is performed. If the RSD3 of the three values meets the criterion and the mean of the three results falls within the content limits, the sample passes. This can be repeated up to a maximum of six determinations. The sample can only be rejected if the mean is outside the content limits and the criterion for the RSDn is met. If at any stage the RSDn is greater than the value listed in Table 2, further determinations are useless because it can be predicted that the RSD will not meet the criterion. Instead, the reason for the poor repeatability should be investigated. As a consequence, the sample can neither be accepted nor rejected.
进行2次检测，如果2次检测RSD小于2次检测的允许最大RSD值（见表1），且平均值在含量限度内，则样品合格。如果两个条件其中任意一个不满足，则应再加测一次。如果3次结果RSD值符合标准，且3次结果平均值在含量限度内，样品合格。这种方法可以一直重复直到得到6次检测结果。只有当平均值不在含量限度内，但RSD符合要求时，样品可以判定不合格。如果在任何一步RSD值大于表2中的值，则加测没有用，因为可以预见RSD是不会符合标准的。这时，应调查找出重复性差的原因。这时，对样品不能判断合格或不合格。只有在含量检测过程中发现错误则应判定原检测结果无效，重新开始含量检测。
This approach is illustrated in Figure 1. 本方法在图1中以流程图表示。

 图1 – 方法1的决策树

Table 1 — If the RSDn is greater than the values listed, additional assays are required up to a
maximum of 6.表1 - 如果RSDn比下表数值更大，需要进行最多六次含量检测

B
Number of determinations 检测次数(n)
2
3
4
5
6
1.0
0.11
0.29
0.42
0.52
0.60
1.5
0.17
0.44
0.63
0.78
0.90
2.0
0.22
0.59
0.84
1.04
1.20
2.5
0.28
0.73
1.05
1.30
1.50
3.0
0.33
0.88
1.26
1.55
1.80
5.0
0.56
1.48
2.13
2.62
3.03
10
1.12
2.97
4.26
5.25
6.06
≥15
1.68
4.45
6.38
7.87
9.09
B = Upper specification limit - 100

Table 2 – If the RSDn is greater than the values listed, additional determinations are useless.
表2 – 如果RSDn比以下数值更大，则更多的检测是没有用的
B
Number of determinations 检测次数(n)
2
3
4
5
6
1.0
1.34
0.95
0.77
0.67
0.60
1.5
2.01
1.42
1.16
1.01
0.90
2.0
2.68
1.90
1.55
134
1.20
2.5
3.35
2.37
1.94
1.68
1.50
3.0
4.02
2.85
2.32
2.01
1.80
5.0
6.70
4.70
3.80
3.30
3.00
10
13.40
9.40
7.60
6.60
6.00
≥15
20.10
14.09
11.39
9.87
9.00

Approach 2: Active pharmaceutical ingredient, 3 initial determinations
方法2：原料药，3次原始检测
Perform three determinations. If the RSD3 is smaller than the RSDmax permitted for three determinations (see Table 1), and the mean falls within the content limits, the sample can be accepted. If either of these two conditions is not met, three further determinations are performed (unless the RSD exceeds the value in Table 2, in which case further assays are useless). If the RSD6 of the six values is smaller than the RSDmax permitted for six determinations, and the mean of the six values falls within the content limits, the sample can be accepted. The sample can only be rejected if the mean is outside the content limits and the criterion for the RSD is met. If theRSD is too large, the reason for the poor repeatability should be investigated and, in such circumstances, the sample can neither be accepted nor rejected.
进行3次检测，如果3次结果RSD值小于3次检测允许的最大RSD（参见表1），且平均值在含量限度内，则样品合格。如果两个条件中有一个不能满足，则再进行3次检测（如果RSD大于表2中的值，则进行更多次检测是没有用的）。如果6次结果的RSD值小于6次检测允许的最大RSD值，且6次检测平均值在含量限度范围内，则样品合格。仅当平均值在含量限度范围之外，且RSD符合要求的水平时判定样品不合格。如果RSD值太大，则应进行调查找出重复性差的原因，这时，样品不能判定是否合格。只有当含量检测被发现有错误时，结果判定无效，重新测定含量。
This approach is illustrated in Figure 2 该方法图示如图2。

图2 – 第2种方法的决策树

Approach 3: Impurity tests (e.g. Related substances by HPLC)
方法3：杂质检测 （例如HPLC检测有关物质）
Approach 2 may be adapted to quantitative impurity tests, where an analytical/instrumental response is obtained (e.g. peak area). In that case, RSDmax is 10% when results are found near the Limit of Quantitation (LOQ)[3] or Reporting Threshold[4].
当检测原理是获得分析/仪器响应（例如峰面积）时，方法2可以用于定量杂质检测。在这种情形下，如果结果接近定量限（LOQ）或报告限，则RSD最大为10%，
Perform three determinations. Responses below LOQ or with a signal-to-noise ratio of less than 10 should be disregarded. If the RSD3 is smaller than or equal to 10%, and the mean complies with the specifications, the sample can be accepted. If this requirement is not met, three further determinations are performed. If the RSD6 of the six values is smaller than 10% and the mean of the six values complies with the specifications, the sample can be accepted. The sample can only be rejected if the mean does not comply with the specifications and the criterion for the RSD is met. If the RSD6 is still too large, the reason for the poor repeatability should be investigated and, in such circumstances, the sample can neither be accepted nor rejected.
进行三次检测。低于LOQ或信噪比小于10的响应要忽略。如果3个结果的RSD小于等于10%，平均值符合质量标准，则样品可以接受。如果不符合上述要求，则需要进行另3次检测。如果6个结果的RSD小于10%，且6个结果的平均值符合质量标准，则样品可以接受。只有当平均值不符合质量标准，且RSD不符合标准时，该样品才会被判定不合格。如果6个结果的RSD太大，则应调查重复性差的原因，在这种情形下，样品不能判定不合格，也不能判定为合格。
This approach is illustrated in Figure 3. 本方法见图3.

图3 第3种方法的决策树

[1] Observed repeatability should be interpreted as repeatability between independent sample determinations, either with regards to standard deviation (sd) or relative standard deviation (RSD).  
[2] Note that using RSD2 is equivalent to using the relative range if the requirement for n = 2 is multiplied by . 2
[3] Quantitation limit, as defined in “Validation of Analytical Procedures: Text and Methodology Q2(R1)”, ICH, 1994.  
[4] Reporting threshold, as defined in the European Pharmacopoeia 7.4, Chapter 5.10. Control of impurities in substances for pharmaceutical use”.  

Approach 4: Finished products  
方法4：制剂检测
For recently-registered products with fully validated analytical methods, information regarding repeatability and/or intermediate precision of the test method is supplied in the application file. The repeatability might be reported with different degrees of freedom depending on the experimental design. The minimum degrees of freedom is 5, as given in the ICH guideline.
对于最近注册的产品，其检测方法关于重复性和中间精密度的验证信息在申请文件中提供。重复性可能根据实验设计的不同自由度进行报告。在ICH指南中给出的最小自由度为5。
During the assessment process, the performance characteristics of the quality control procedures are evaluated against the specification limits proposed by the manufacturer. When approved, the results of the tests performed must fall within the specification limits.
在评估过程中，检测方法特性会与生产商提议的质量标准限度进行对照。批准时，检测结果应符合质量标准限度。
When a product is to be tested at an OMCL, the MAH file is consulted in order to find suitable conditions for the test method and also to get information on its performance characteristics. The repeatability of the results obtained during testing can therefore be used as a quality indicator and can be checked against the values given in the dossier. Nevertheless, the OMCL might find that internal quality control criteria for evaluating the repeatability of the results of the test are suitable for the intended purpose.
如果由OMCL进行检测的产品，其MAH文件在进行审阅，则检测应证明其检测方法适用条件，同时获得检验方法表现情况的信息。检测过程中所得的结果的重复性可以用作质量指示，与申报文件中的数值进行对比检查。
The observed standard deviation varies between testing events, following a skewed distribution. To test whether standard deviations (or, rather, variances) are not significantly different, the quotient of two variances is calculated and compared to the critical F-value at a specified probability for the relevant degrees of freedom. In Tables 3 and 4, the critical F-values at the 5% level have been used to calculate the maximum allowable standard deviations under the assumption that the observed repeatability is not significantly worse than that reported in the dossier. If the OMCL chooses to use its own in-house repeatability criteria, the use of the data described in Tables 3 and 4 is not applicable and the comparison should be done with the predefined RSD criteria. This should be taken into account in points 3, 4, 5 and 6 below.
所得的标准偏差可能会受到不同影响。如果标准偏差和变动性都没有显著不同，则计算两个变异值的商，与对应的自由度下的指定可能性的关键F值比较。在表3至表6中，5%水平的关键F值用于计算允许最大标准偏差，前提是假设观察到的重复性未显著差于文件报告值。
1. Find the RSD and the degrees of freedom for the repeatability in the dossier.
计算RSD，文件中重复性的自由度
2. Perform three determinations and obtain the results in % of the label claim. Calculate the mean and the relative standard deviation.
进行3次检测，得到标签所声称的百分比结果。计算平均值和RSD。
3. Check in the relevant Table 3 or Table 4 if the RSD obtained for the repeatability is larger than the critical va