The use of platelet-rich plasma (PRP) in orthopaedic practice has grown exponentially over the last decade.1–3 PRP can be defined as any autologous blood preparation in which the platelets have been concentrated to levels exceeding that in the whole blood from the same patient. These therapies aim to deliver proregenerative growth factors (GFs) and cytokines, which are released from a concentrated pool of degranulating platelets, to the site of pathology. GFs released by platelets have been demonstrated to perform proregenerative functions in vitro, including promoting stem and progenitor cell proliferation and recruitment, modulating inflammatory responses, and stimulating angiogenesis.4,5 The rationale for using platelets is to augment or accelerate healing in many musculoskeletal conditions. The autologous nature, favourable safety profile, and simplicity of production of PRP makes it an appealing approach to treatment. PRP preparations vary considerably, and the optimal preparation for treating different musculoskeletal conditions remains unknown. Overall, the clinical use of PRP has greatly outpaced the evidence supporting its application.5,6

The heterogeneity of the processing methods used to prepare PRP, as well as the lack of reporting of even basic characteristics and compositions of these preparations, is a key barrier to understanding PRP’s clinical effects. The vast majority of clinical studies evaluating PRP preparations have not provided sufficient information to allow interpretation or replication of protocols.7 This makes interpretation of outcomes difficult and makes comparison between studies almost impossible.

At present, there is no all-encompassing and universally accepted system to allow classification of PRP and other autologous blood preparations. An ideal classification should be simple to use, should be reproducible, and should focus on characteristics that are relevant to the prognosis and therapeutic decision-making. Although numerous classification systems for PRP formulations have been proposed, none has achieved universal acceptance.8–13 In this annotation, we outline existing systems used to classify preparations of PRP, highlighting the need for a standardized universal nomenclature and classification system for blood-derived products.

Dohan Ehrenfest classification (2009): with appreciation that different components of PRP preparation may influence therapeutic effect, Dohan Ehrenfest et al8 proposed a classification that separated products using two basic parameters: the presence or absence of leucocytes, and the fibrin architecture (Table I).8–13 This separation resulted in four main PRP subtypes. 1) Pure PRP, or leucocyte-poor PRP: preparations with no or low leucocyte levels and with a low-density fibrin network after activation. The PRP products in this group can be used as liquid solutions or in an activated gel form. 2) Leucocyte-rich and PRP: preparations with elevated values of leucocytes and with a low-density fibrin network after activation. The PRP products in this group can also be used as liquid solutions or in an activated gel form. 3) Pure platelet-rich fibrin, or leucocyte-poor platelet-rich fibrin: preparations with no or low leucocyte levels and with a high-density fibrin network. Products from this group only exist in a strongly activated gel form and cannot be injected as a liquid solution. 4) Leucocyte-rich and platelet-rich fibrin: preparations with elevated values of leucocytes and with a high-density fibrin network. They only exist in a strongly activated gel form. Although this classification is simple, quantitative values are lacking.

PAW classification (2012): DeLong et al9 described the platelet, activation, white blood cells (WBCs), or ‘PAW’, system, which is based on three components: 1) the absolute number of platelets; 2) the manner in which platelet activation occurs; and 3) the presence or absence of WBCs (Table I). Additionally, the importance of precise determination of neutrophil levels was emphasized, and the authors incorporated a subcategory for neutrophil count. The three main parameters were described as: 1) platelet concentration, measured in platelets per millilitre and categorized as P1 (≤ baseline), P2 (> baseline to 750 000 cells/μl), P3 (> 750 000 to 1 250 000 cells/μl), and P4 (> 1 250 000 cells/μl); 2) total WBC content, identified as either above or below/equal to baseline levels, with α (above) added if neutrophils were included in the buffy coat or β (below) added if neutrophils were filtered out; and 3) activation method. Endogenous activation was not given a designation. However, if an exogenous external activator was used, it was documented with an ‘X’. This classification ensures a more accurate and definitive reporting of the concentration of platelets (number of platelets per millilitre) but leucocytes remain described as a binary measure (above or below/equal to baseline levels).

Mishra’s classification (2012): Similarly to DeLong et al9, Mishra et al10 also published a classification based in platelet concentration, the presence or absence of WBCs in the PRP, and the use of agonists to activate PRP (Table I). However, the authors classified the variables in a different way to DeLong et al,9 identifying four types of PRP: 1) type 1 contains an increased concentration of platelets and WBCs over baseline, and is not activated by an exogenous activator such as thrombin or calcium; 2) type 2 contains both increased platelets and WBCs and is activated with thrombin and or calcium; 3) type 3 contains only an increased concentration of platelets without any WBCs and is not activated prior to application; and 4) type 4 contains only an increased platelet concentration and is activated with thrombin and/or calcium. Subtype A contains an increased platelet concentration at or above five times baseline. Subtype B contains an increased platelet concentration less than five times baseline. Nevertheless, the three preceding classifications do not take into account the final volume of the preparation delivered.

In 2015, Mautner et al11 highlighted the potential detrimental effects of red blood cells (RBCs) on PRP activity due to their chondrotoxic and proinflammatory effects, recommending that the presence or absence of RBCs in PRP preparations should be reported. Consequently, they established a PRP classification system called platelet, leucocyte, RBCs, and activation (PLRA) classification (Table I). The authors emphasized the importance of describing the platelet count (absolute number/μl), leucocyte content (as positive or negative), percentage of neutrophils, RBC content (as positive or negative), and activation (yes or no for exogenous activation).

DEPA classification (2016): In 2016, Magalon et al12 introduced two new concepts that had not been addressed by the previous classifications. First, the authors stressed the importance of the efficiency of production of the PRP, which corresponds to the proportion of the platelets recovered in the PRP from the blood. Second, the authors emphasized the importance of reporting the purity of the obtained PRP. They defined purity as the relative composition of platelets, leucocytes, and RBCs in the obtained PRP. The Magalon classification is called dose of platelet, efficiency, purity, and activation (DEPA; Table I). The four main parameters are described as follows: 1) dose is calculated by multiplying the number of platelets in PRP by the obtained volume of PRP, classifying from A (> 5 billion platelets) to D (< 1 billion platelets); 2) efficiency corresponds to the proportion of platelet recovery from whole blood, ranging from A (> 90%) to D (< 30%); 3) purity corresponds to the proportion of platelets in the PRP compared with RBC and leucocytes varying from A (> 90% platelets in relation to other cell types) to D (< 30%); and 4) activation related to exogenous activation.

MARSPILL Classification (2017): Finally, Lana et al13 established a PRP classification system called MARSPILL (method, activation, red blood cells, spin, platelets, image guidance, leucocytes, and light activation; Table I). In this system, the authors recommend taking into account four variables that they considered relevant that were not included in previous classifications. Specifically, they propose to include: 1) whether the PRP is prepared in an automated manner or manually; 2) the number of spins that are performed during PRP preparation; 3) whether image guidance is used in order to identify the correct site of application; and 4) if photoactivation was applied to the PRP.

Therefore, the parameters in this classification are described as follows. 1) Method: whether preparation was automated, by machine (M), or handmade (H). 2) Activation: whether PRP is activated (A+) or not (A-). 3) RBCs: rich (RBC-R) or poor (RBC-P), with poor classified as a reduction of approximately 15-fold compared with baseline. 4) Spin: whether one (SP1) or two (SP2) spins are used during the PRP preparation. 5) Platelet concentration: categories of PL two- to three-fold, PL four- to six-fold, PL six- to eight-fold, and PL eight- to ten-fold over baseline. 6) Image guided: guided (G+) or not guided (G-). 7) Leucocytes: leucocyte-rich PRP (Lc-R) or leucocyte poor- PRP (Lc-P) depending on whether the concentration of leucocytes is higher or lower than baseline value, respectively. 8) Light activation: light activated (L+) or not (L-).

However, the relevance of image guidance and light activation are unknown. It is acknowledged that placement in a certain anatomical location (e.g. hip) would be dependent on image guidance, and limited evidence for photoactivation of PRP exists for the nonsurgical management of osteoarthritis of the knee.14

There are a number of potential explanations why none of these systems have achieved universal acceptance or widespread use. The complexity and inherent variability of biological preparations of PRPs precludes a comprehensive listing of all constituents within a classification system or communication tool. However, researchers or clinicians may feel that the current systems do not account for variability in preparations described, nor do they represent the raw data variables from blood components. As such, they do not wish to categorize their own PRP preparations together with previously published preparations that may have produced unfavourable results. In addition to PRP, there are a growing number of autologous blood preparations being proposed for musculoskeletal applications that do not contain concentrated platelets.15 These include platelet lysate, autologous protein solution, autologous conditioned serum, and platelet-poor plasma. It is crucial for future classification systems to encompass all autologous blood preparations.

There are logistical and financial barriers due to the cost and complexity associated with cell and biochemical analysis that might explain why some variables have not been included in the classifications systems. Nonetheless, a comprehensive and granularly structured system is required to allow classification based on similar predefined characteristics. Without such a system, access to information in an efficient and timely manner is limited, which jeopardizes the advancement of the field.

Blood-derived products, including PRP, are now widely used to treat a range of musculoskeletal pathologies, despite a lack of robust evidence to support efficacy.1,16–23 This popularity has been fuelled by a lack of efficacious non-surgical options to treat some musculoskeletal pathologies combined with the ease of use, perceived safety, and commercial enthusiasm.2,24 In response to growing clinical use of biological therapies, the American Academy of Orthopaedic Surgeons (AAOS) recently convened a collaborative symposium that aimed to identify strategies to facilitate high-quality research into autologous products and encourage the responsible and evidence-based use of these therapies.2

Numerous basic science and animal studies support the notion that targeted preparations of PRP may have favourable effects on the healing process of various musculoskeletal tissue types.25–29 Given these positive findings, it is easy to understand the interest in treatment with PRP. However, in the rush to clinical translation, numerous clinical trials have been performed without full characterization of PRP attributes nor optimization of preparations.5–7,30 Even though the basic science data supporting the potential beneficial effects of GFs in augmenting connective tissue healing are promising, the clinical benefits of using PRP have not been universally achieved.5,6 A critical analysis of the PRP literature suggests two primary reasons for the absence of alignment between the expectation derived from basic science and clinical reality. One reason may be the heterogeneity in the PRP preparation protocols and the final composition of the PRP delivered. Variations in the volume of whole blood taken, the platelet recovery efficacy, the final volume of plasma in which the platelets are suspended, the presence or absence of WBCs, and the addition of exogenous activators to induce fibrin formation can all affect the character and potential efficacy of the final PRP product.31–37 When analyzing the PRP preparation protocols and PRP composition utilized in clinical trials for the treatment of musculoskeletal diseases, only 10% of the studies provided comprehensive reporting that included a clear description of the preparation protocol that could be used to repeat the method by subsequent investigators.7 Furthermore, only 16% of the studies provided quantitative metrics on the composition of the PRP final product. Even within a given PRP separation technique, a high degree of inter- and intrasubject variability in the composition of PRP exists, which must be addressed for improved reporting.38 This will contribute to the inconsistency of results reported in the literature. It is essential that a precise and stepwise description of the PRP preparation protocol is provided to allow comparison among studies and enable reproducibility.

Nevertheless, even if the PRP preparation protocol is clearly described, and if detailed quantitative metrics on the composition of the final product is provided, there is an additional key challenge, namely to communicate autologous blood preparations effectively using standardized nomenclature. There is a wide spectrum of PRP preparation protocols and formulations used in the different studies, all grouped under the term ‘PRP’.3,16–23Therefore, the success or failure of a specific PRP product is not necessarily predictive of all PRP preparations. As shown, in the last six years, five different classifications were described.9–13 To our knowledge, no evidence-based classification guidelines specifying the optimal PRP preparation have been reported for different musculoskeletal disorders. A classification that accurately characterizes the specific preparation of PRP will enable correlation with validated outcome assessment tools in studies exploring specific indications.

A future consensus classification on PRP should incorporate complete information from the final PRP delivered to allow standardized comparison between studies. Such a classification will have to incorporate a quantitative approach to reporting on all main components of PRP while staying simple and practical. Therefore, the authors feel that three essential factors that need to be represented are: 1) platelets; 2) WBCs and percentage of neutrophils; and 3) RBCs. Furthermore, concentration and dose of PRP has to be recorded to document total number of the components delivered.

We believe that platelet number should be reported as platelet concentration (platelet number per millilitre), along with the volume of PRP delivered. We do not recommend reporting the platelet concentration as a multiple of the baseline because the baseline platelet count can vary significantly between patients.

WBCs should also be reported as concentration, and WBC differential should be reported when possible. Leucocyte concentrations have a strong influence on the growth factor and cytokines delivered to the target tissue.39 Given the multifunctional roles of WBCs, it is possible that WBCs levels in PRP, or specific WBC subtypes, may be beneficial in specific musculoskeletal conditions (such as chronic tendinopathy),40 while being detrimental in other others (such as osteoarthritis or acute muscle strain).17,41 It is likely that the need to include WBCs in the PRP preparation will vary by indication, and the WBC concentration should be documented. We believe that only characterizing WBCs as above or below the baseline is insufficient. The concentration of leucocytes above the baseline can vary markedly and this will probably influence clinical outcomes.39 Regarding WBC differential, and especially neutrophils, we believe that they should be measured and reported as a percentage of total WBCs.

Concerning RBCs, basic science studies have showed that they can adversely affect platelet function by altering local pH and promoting inflammation.37,42 Furthermore, blood-derived proinflammatory cytokines trigger chondrocytes and induce the production of cartilage-degrading proteases causing chondrocyte death. Additional clinical studies are needed to determine the clinical effects of different concentrations of RBCs on inflammation and wound healing. In the meantime, we recommend that the dose of RBCs should be documented in the same way as platelets and WBCs.

Most authors have included activation in their classifications.10–13 Platelet activation, and subsequent release of cytokines and GFs, can be initiated by a number of methods such as shear forces caused by fluid flow, contact with a variety of materials including fibrillar collagen and basement membranes of cells, and thrombin.27,43 Once the PRP is activated, a fibrin network will begin to form and plasma will begin to solidify to create a fibrin clot or membrane. Some PRP preparation techniques advocate the activation of platelets, with thrombin and/or calcium chloride (CaCl2), before PRP delivery to guarantee that the bioactive factors contained in the α granules will be secreted and readily available. In contrast, opponents of using activators suggest that natural activation via interaction with the patient’s own collagen is a superior route.44 When activated, platelets begin secreting their GFs immediately.45 Approximately 70% of these GFs are secreted within the first ten minutes after activation; within an hour almost 100% have been secreted.45 Although activation changes the properties of PRP, the biological consequences of exogenous activation of PRP on tissue healing have yet to be delineated. We recommend that if PRP is activated exogenously, this should be part of the classification of the final PRP product.

Finally, guidance on the reporting of relevant parameters describing preparation protocol of PRP might also need to be included in the methods section of every study to allow reproducibility. Some of the parameters that might be suggested includes type of anticoagulant, preparation technique (including spin rate and/or g-forces and duration), and make and model of the centrifuge.

Take home message

– At present, there is no all-encompassing and universally accepted system to allow classification of platelet-rich plasma (PRP) and other autologous blood preparations.

– An ideal classification should be simple to use, should be reproducible, and should focus on characteristics that are relevant to the prognosis and therapeutic decision making.

– A future consensus classification on PRP should incorporate complete information from the final PRP delivered to allow standardized comparison between studies. Such a classification will have to incorporate a quantitative approach to reporting: 1) platelets; 2) white blood cells and percentage of neutrophils; and 3) red blood cells. Furthermore, concentration and dose of PRP has to be recorded to document total number of the components delivered.