Issues in stability of drugs for infusion

The need to have real stability data for drugs, and especially for anticancer molecules in ‘in-use’ situations is critical. Indeed, post-dilution or reconstitution stability data are frequently limited to 24 hours only for bacteriological reasons, considering that these practices are performed under uncontrolled conditions on the ward, and that the drug will be administered quite quickly thereafter. It is a basic application of the care principle, sometime justified by a potential, but rare, risk of observing leachables during longer conservation. However, if no specific justification is furnished by the manufacturer, this dogmatic position cannot be sustained further because, in many countries, reconstitution and preparation of anticancer drugs takes place in pharmacy-based centralised compounding units, in a controlled and validated environment with expert staff. Therefore, the stability limits do not depend on possible contamination but the only relevant issue is the actual chemical and physical stability, which could potentially be much longer. As an example, it has been demonstrated that the monoclonal antibody, rituximab, diluted in polyolefin bags and kept at 4°C or 22°C, is stable for more than six months despite available data in the package insert or in the Summary of Product Characteristics (SPC) indicating a stability limit of 24 hours.¹

A comparable extended stability has been also demonstrated for the lyophilised antibody, trastuzumab, after reconstitution and storage in plastic bags.² Interestingly, and ironically, this discrepancy is indirectly acknowledged by the manufacturer. Indeed, the stability of ‘multi-use’ trastuzumab reconstituted by bacteriostatic water (formulation not available in Europe) is claimed to be 28 days in the SPC, but is only 48 hours for the same product reconstituted with water for injection. This fact demonstrates that the latter stability limit is based only on possible risk of biological contamination and not on true physicochemical stability.³ However, the strict application of the SPC recommendation by European pharmacists having no access to the multi-use formulation may induce a notable loss of this costly drug in cases of delayed use, although the drug still remains stable. Under current economic constraints, this kind of wastage is obviously unacceptable.

Moreover, the real physicochemical stability of a drug in simple situations, such as stability limit at 4°C in a polyolefin bag, may be insufficient in many practical cases that might be observed in daily practice. As an example, some hospitals pharmacists would like to use pneumatic conveying systems to deliver the bags from the centralised unit to the various wards, in order to avoid delays. Therefore, the problem is what happens during the stress induced by the pneumatic system for antibodies that are considered to be highly sensitive to mechanical stresses, such as shaking?

Obviously, centralised preparation units are faced with new needs and/or daily in-use situations, which include:

All of these situations require the pharmacist to use a set of relevant stability data to highlight the professional decision.

However, stability studies performed by the pharmaceutical industry are predominantly designed to fulfil licensing requirements. Meanwhile, several generic companies have the results of their extended stability studies also listed in the SPCs.

When originator medicines are being licensed, little attention is given to the practical use of these drugs and there is no clear awareness that pharmaceuticals start a new life once they are prepared for patient administration. It must be also underlined that, for a new drug, stability studies have been performed several years before the marketing authorisation, in a conventional way by preclinical formulation and analytical pharmaceutical scientists, without knowledge, and consideration, of the reality of the in-use situations and/or evolution of clinical practices. Therefore, most relevant stability data cannot be furnished by the manufacturer, regardless of possible financial interests to maintain short-term stability limits for very costly products, especially original drugs in monopolistic situations.

To offer the pharmacy community some industry-independent arguments for a better use of drugs, several academic teams are implicated in the in-use stability studies, especially for anticancer drugs.

Chemotherapy agents are generally considered as having a very narrow therapeutic range and are very toxic themselves, although this is not always strictly true. Generally speaking, stability is the ability of a drug to retain its physical, chemical, microbiological and biological properties within previously defined, specified limits. However, the concept of in-use stability is more extensive, referring to the stability of a drug not only determined under conventional situations, but also taking into account variations observed in clinical practice.

For drugs, the general rule applied is that a drug remains stable in clinical practice (that is, at recommended dilution and vehicle) until 90% of its initial value (T90 value). Thus, this 10% degradation as a stability limit has been widely used in published stability studies. However, for anticancer drugs, some authors considered that in-use stability limits should be defined through a product-by-product approach considering the afferent therapeutic index, clinical use, safety and potency, the pharmacodynamic/phamacokinetic variability, and the total cumulative dose. Indeed, depending on the drug, this limit could lead to unacceptable toxicity and/or loss of efficiency.^4,5

As an example, for the same dose of 5-FU by infusion, the area under the curve between patients can vary by more than 500%.⁶ Thus, the administration of only 90% of the theoretical amount of 5-FU will be clinically irrelevant in terms of efficacy. However, increased risks associated with its degradation products (DPs), such as fluoroacetic acid, even at low levels, must also be considered. Thus, the stability limit for 5-FU will depend on the degradation pathways. If non-toxic DPs were only observed, then a 10% degradation limit will be acceptable. By contrast, if toxic DPs are identified and published in the literature, the stability limits must be less than 5% of degradation.

Consequently, a careful study of DPs is of major importance in stability studies for anticancer drugs. However, few studies on the toxicity of DPs are available. As an example, it has been suggested that the increased cardiac toxicity in patients receiving  high doses of 5-FU could be caused by small quantities of fluoromalonaldehyde and fluoroacetaldehyde DPs resulting  from storage in basic medium.⁷

Finally, a recent European consensus recommends, as a general rule, a stability limit of 95% for anticancer drugs.

Physical stability of anticancer drugs in infusion is largely neglected in many in-use stability studies. Physical instability means colour change, opalescence, turbidity, precipitation, or formation of sub-visible aggregates. Physical instability is a major challenge for biologics because the first instability sign of many proteins submitted to a stressor (thermal, mechanical, light) is aggregation.⁵ Therefore, a well-executed, in-use stability study should include careful examination of physical instability.

If in-use stability studies are not available from the manufacturer, those performed by academic teams must follow strict methodologies in terms of protocol design, analytical methods and statistical analysis of results. This point is of paramount importance because manufacturers’ non-based stability data implicate that its responsibility cannot be engaged. Pharmacists who want to use independent stability data in daily practice have to be sure that these data are correct, relevant and standardised/validated. The problem is that much of the published stability data are not of the required quality and can be easily criticised.

Classically, low-level publications use an insufficiently validated method, for example without verifying that the drug chromatographic peak did not contain a degradation product. Thus, the method used to analyse the drug must be proved as indicative of stability as detailed in the ICH guideline Q5C and other recommendations,^8,9 and as used in good quality papers.^10,11 Another low scientific approach is to use a biological method to estimate the stability of drugs, such as monoclonal antibodies, enzymes, antifungals or antibiotics. This approach can be only considered as a complementary determination among a set of analytical stability-indicating methods permitting the unambiguous separation of the initial drugs from its degradation products. As an example of bad methodology, a recent claim that two monoclonal antibodies (cetuximab and pamitumumab) were stable in polyolefin bags was only based on an ELISA test, without confirmation by complementary methods demonstrating the stability of the antibody structure.¹²

Indeed, a degradation product may also have pharmacological activity, which cannot be distinguished from the parent drug in a biological assay. Moreover, subtle modifications of the secondary and tertiary structures must be analysed (and characterised if these occur) during any in-use stability study of a protein, especially monoclonal antibodies, which are increasingly being used in oncology.^1,2,5

Beside general books and reviews on the stability of drugs,^13,14 in the particular case of anticancer drugs, a recent European conference consensus has defined general rules to perform good quality stability studies.^5,8 Detailed and tabulated in-use stability data for anticancer drugs are available in the literature.¹⁵ Moreover, the well-known and free access Stabilis® database, provides, in more than 28 languages, up-to-date information on the stability of several hundred drugs, with references and an invaluable evaluation of the level of proof.¹⁶ However, quite often industry is asked for a recommendation for legal reasons.

In-use stability data are a vital tool for hospital pharmacists to optimise their practices. These studies must be strongly encouraged but the price to pay is a rigorous scientific approach, especially in terms of analytical methods. Shelf-life specification is the only measure that should be binding for industry. Because many of the stability studies cannot be achieved by drug manufacturers, regardless of good or bad reasons not to perform them, independent studies from academic and hospital teams must be developed. However, the problem of financing these studies is an issue that will need to be resolved.