**1. Introduction**

It is common in biomedical research for protocols to require blood collection from mice (Mus musculus) in order to measure a range of circulating products. The small size of these animals makes such procedures challenging, but a range of sampling methods are documented and widely used. Common considerations in the selection of sample techniques are their practicality and ease of use, the ability to attain the desired blood volume, sample quality and impact on animal well-being [1,2].

Retrobulbar bleeding (retroorbital) enables acquisition of larger blood volumes (e.g., 0.2–1 mL) [3], but has been controversial due to the risk of substantial tissue damage to the eye [3,4]. Anecdotally, it appears that this technique has fallen out of favour, particularly in some jurisdictions, such as Australia. This has led to the development of alternative methods. The most common alternative is facial (also commonly called submandibular) vein venepuncture [5]. Despite the rise in popularity of this method, perhaps driven by the aesthetically repugnan<sup>t</sup> use of the ocular area, veins in the facial region lie beneath other important tissues such as glandular tissue [6]. This method then also poses a risk of causing secondary complications via tissue damage. Sublingual vein puncture is another alternative method which yields large-volume collections [4]. In contrast to the submandibular technique, sublingual sampling is generally performed under anaesthesia to immobilise the animal [4]. Anaesthesia, as an adjunct, has the potential to impact both positively and negatively on animal well-being, through minimisation of tissue damage [1], or 'hangover' e ffects from the drugs. Anaesthetic use will also influence the practicality of the technique due to equipment needed and the time taken to perform.

A range of methods are available for the attainment of smaller volume samples (approximately 0.1–0.15 mL), or for frequent repeat sampling. The tail is commonly used as a blood collection site. A range of collection techniques are described, including targeted collection from the lateral tail vein [7,8], tail tip amputation [9–11], and tail incision through cut of the veins [12,13]. Anaesthesia has been regarded as unnecessary for this site, enabling multiple repeat samples. However, warming of the animal may be required to encourage vasodilation [3]. This may add to overall impact on animal well-being. The lateral tarsal or saphenous vein is a common alternative to the tail. Similarly, it requires no anaesthetic and has the added advantage of not requiring external methods for vasodilation. Removal of the scab enables serial blood sampling [3].

Whilst pain, discomfort and physiological stress arising from blood sampling are likely to be short-lived, as one of the most common procedures performed on laboratory animals, researchers and animal ethics committees have a duty to utilise or promote the method with least impact on animal well-being. Furthermore, with the demise of retrobulbar sampling on supposed ethical grounds, it is imperative that an evidence-based approach to the selection of alternative methods is used. Whilst there have been a range of studies investigating the impact of a number of the sampling techniques on mice well-being, these studies typically only contrast a few techniques, and are practically limited in terms of sample sizes. The aim of this systematic review is to present the evidence related to common recovery blood sampling techniques in mice, with regard to animal well-being. Through identification of all relevant evidence, assimilation of study findings to increase statistical power, and study appraisal, it is our intention that this systematic review will provide increased strength of evidence to better inform researchers, ethics committees, and policy makers in their decision making.

### **2. Materials and Methods**

A priori protocol was created for this review and has been registered on the SYstematic Review Centre for Laboratory animal Experimentation (SYRCLE) database for animal intervention studies [14].
