The Impact of Systemic Treatment With Flunixin Meglumine on the Two Antiproteases Alpha‐2‐Macroglobulin (A2M) and Alpha‐1‐Antitrypsin (A1AT) in Equine Plasma and Platelet‐Rich Fibrin (PRF)

ABSTRACT

Purpose

To evaluate the effect of flunixin meglumine, a systemic non‐steroidal anti‐inflammatory drug (NSAID), on the antiproteases, A2M and A1AT, in equine plasma and PRF.

Methods

A prospective cross‐sectional study. Plasma, solid PRF, and liquid PRF were collected from a control‐group that included healthy horses not treated with NSAIDs within the last 3 months. The NSAID‐group were ophthalmology patients with no systemic inflammatory disease, and all received a full dose of flunixin meglumine (1.1 mg/kg BID PO or IV) for three or more days prior to collection. Samples were collected and prepared using a standardized protocol for plasma and PRF preparation. Concentrations of A2M and A1AT were analyzed with ELISA kits. Statistical analysis with Mann–Whitney U tests and Wilcoxon signed‐rank tests was performed using significance set at p < 0.05.

Results

No significant differences were found between the control‐group (n = 12) and NSAID‐group (n = 12) for any of the blood products (p > 0.05). Plasma had a significantly higher amount of A2M compared to solid and liquid PRF (both p < 0.001). Solid PRF had the highest concentration of A1AT and was significantly higher when compared to liquid PRF (p < 0.001), but not when compared to plasma (p = 0.071).

Conclusions

The systemic NSAID, flunixin meglumine, does not appear to significantly alter antiprotease concentrations in equine plasma or PRF. Although PRF should not be used as a replacement for plasma, solid PRF may be a good adjunctive therapy for the treatment of keratomalacia due to its high A1AT concentration.

1. Introduction

The healthy equine cornea provides an appropriate barrier to the invasion of bacteria and fungi [1]. However, once this barrier is disrupted following an injury, pathogens can evade the innate defense mechanisms of the equine cornea and gain access to the stroma. In cases of stromal ulcerative keratitis, destruction of the corneal stroma is caused by uncontrolled protease activity. These proteolytic enzymes are normally in balance with natural protease inhibitors in the preocular tear film [1, 2]. However, microorganisms, leukocytes, corneal epithelial cells, and fibroblasts all produce excessive amounts of proteolytic enzymes upon damage to the cornea [1, 2, 3, 4]. Two important proteolytic enzymes for the equine cornea are matrix metalloproteinases (MMPs) and serine proteases such as neutrophil elastase (NE) [1, 3, 4, 5]. Leukocytes produce NE in higher concentrations in cases of ulcerative keratitis [3]. Concurrently, studies have shown that both MMP‐2 and MMP‐9 concentrations increase considerably leading to further collagenolysis, corneal liquefaction, and keratomalacia [3, 6, 7]. If uncontrolled, these cases can rapidly progress to corneal perforation [1].

Treatment of stromal ulcerative keratitis and keratomalacia involves targeted antimicrobials, controlling uveitis, and reducing proteolytic activity. Studies have shown that persistent and high levels of MMP‐2 and MMP‐9 activity in the tear film correlate with progressive corneal damage [8]. Thus, controlling proteolysis is paramount to improving clinical signs, enhancing epithelial healing, and preventing ulcer progression [6, 7, 8]. A common therapeutic modality used is equine autologous serum or plasma, which provides the two antiproteases alpha‐2‐macroglobulin (A2M) and alpha‐1‐antitrypsin (A1AT) directly to the corneal surface. Equine serum has been shown to decrease corneal proteolytic activity by 90% in vitro [7]. Another component of treating stromal ulcerative keratitis is controlling secondary uveitis. Systemic flunixin meglumine is often the treatment of choice for reducing intraocular inflammation and iridocyclitis [1]. As a non‐steroidal anti‐inflammatory medication, flunixin meglumine inhibits the enzyme cyclooxygenase thereby reducing the synthesis of prostaglandins that mediate inflammation and has been shown to modulate the acute phase response and indirectly reduce serum amyloid A (SAA), a positive acute phase protein [9]. However, despite its widespread use in horses with stromal ulcerative keratitis, no studies have evaluated its effects on antiproteases which are also positive acute phase proteins [10].

Additionally, alternative or adjunctive therapies are being investigated for their potential use in ulcerative keratitis such as platelet‐rich fibrin (PRF). PRF is an autologous second‐generation platelet concentrate that contains high concentrations of growth factors and cytokines [11, 12, 13]. PRF has been shown to be a safe therapeutic option to promote corneal ulcer healing [11, 14, 15]. However, no studies have evaluated PRF from horses for its potential antiprotease properties.

The purpose of this study was to evaluate which effect systemic flunixin meglumine has on the two antiproteases, A2M and A1AT, in equine plasma as well as in solid and liquid PRF. This study will be useful in determining if the quality of autologous plasma and PRF from horses is compromised by systemic anti‐inflammatory medication. Additionally, the second aim of the study will be to compare the concentrations of A2M and A1AT in plasma, solid PRF, and liquid PRF to explore the potential role of PRF in the treatment of keratomalacia.

2. Materials and Methods

This was a prospective cross‐sectional study. Plasma, solid PRF, and liquid PRF samples were collected from two different horse groups, with an appropriate temperament for a one‐time blood collection. The “control‐group” included healthy North Carolina State University (NCSU) research and teaching horses not treated with a systemic anti‐inflammatory medication within the last 3 months. Horses in the “NSAID‐group” were clinical patients at NCSU Equine Hospital's ophthalmology service and were treated with a full dose of flunixin meglumine (1.1 mg/kg PO or IV BID) for three or more days prior to sample collection (Table 1). Flunixin meglumine was administered primarily orally, although some horses received intermittent intravenous doses when an intravenous catheter was already in place prior to a procedure. No horse received flunixin meglumine exclusively by the intravenous route. The study was approved by North Carolina State University's Institutional Animal Care and Use Committee (IACUC protocol #24‐348‐03) and owners of client‐owned horses had signed a client consent form before enrollment.

TABLE 1.

The table gives an overview of the 12 horses enrolled as the NSAID‐group; age, sex (G: Gelding, M: Mare), breed, eye (OD: Oculus dexter, OS: Oculus sinister, OU: Oculus unitas), diagnosis and infectious organism found on microbiology culture, number of full doses of flunixin meglumine (1.1 mg/kg BID PO or IV) given before blood product samples were collected, concentration (ng/mL) of flunixin meglumine measured in plasma at sample collection time.

Horse #	Age	Sex	Breed	Eye: Diagnosis/infectious organism	Flunixin meglumine doses in hospital
1	7	G	Thoroughbred	OS: Ulcerative keratitis/ Candida albicans	7
2	8	M	Thoroughbred	OD: Fungal abscess/Fusarium solani	9
3	20	M	Quarter Horse	OD: Ulcerative keratitis/ Streptococcus equi zooepidemicus	6
4	12	G	Thoroughbred	OS: Ulcerative keratitis/Fusarium solani	6
5	15	G	Thoroughbred	OS: Ulcerative keratitis/infection not confirmed	11
6	9	M	Arabian	OU: Ulcerative keratitis/OD: Staphylococcus equorum ; OS: Infection not confirmed	6
7	16	M	Connemara Pony	OD: Ulcerative keratitis with perforation/infection not confirmed	16
8	4	G	Warmblood	OS: Ulcerative keratitis and primary uveitis/infection not confirmed	6
9	15	G	Thoroughbred	OS: Suspected fungal abscess OS/infection not confirmed	6
10	13	M	Saddlebred	OD: Ulcerative keratitis/Aspergillus fumigatus	6
11	15	G	Quarter Horse	OD: Suspected fungal abscess/infection not confirmed	6
12	23	M	Oldenburg	OS: Ulcerative keratitis/Exserohilum rostratum	6

The plasma and PRF samples were obtained with a vacutainer and a 22‐gauge needle, and blood was collected from either their left or right jugular vein following a quick scrub with 1:50 diluted betadine solution at the area of collection. For the plasma sample, 10 mL of blood was collected into an EDTA tube and centrifuged for 12 min at 4000 RCF before 1.5 mL of the plasma was stored in sterile Eppendorf tubes. The solid and liquid PRF were obtained, prepared, and stored by following the protocol previously described for horses by our research team [11]. For the solid PRF sample, a piece of the solid PRF clot was placed in a sterile Eppendorf tube, and for liquid PRF samples, 1 mL of the buffy coat was collected and stored in a sterile Eppendorf tube. All samples (plasma, solid PRF, and liquid PRF) were stored in Eppendorf tubes at −80°C until analysis.

The biochemical analysis was performed by the University of North Carolina at Chapel Hill's Biobehavioral Laboratory by following the same protocol for how to prepare plasma, solid PRF, and liquid PRF for enzyme‐linked immunosorbent assay (ELISA) analysis, also previously described by our research team [11]. The concentration for A2M and A1AT (pg/mL) were analyzed with ELISA by following the manufacturers' instructions; A2M, mg/dL, Catalog #MBS014509 and A1AT, ng/mL, Catalog #AFG‐GNS‐1640 (MilliporeSigma, Burlington, MA, USA) [13]. The use of these ELISA kits with plasma or other biological fluids was validated by the manufacturer and listed as appropriate sample types. Various anticoagulants, including ethylenediaminetetraacetic acid (EDTA), were considered acceptable anticoagulants for plasma collection, with no specific recommendation to avoid use [16, 17].

Data distributions were examined graphically; the strength of the relationships between products were examined using intraclass correlation coefficients (ICC), as they were considered different measures of the same outcome. Data was expected not to be normally distributed, and descriptive data was presented as medians and range [min/max]. Non‐parametric Wilcoxon signed rank tests were used to evaluate concentration differences between the three blood products (plasma, solid PRF, and liquid PRF). A non‐parametric analysis was used to evaluate differences between the two treatment groups, control‐group versus NSAID‐group, for all blood products and Mann–Whitney U tests were used for these comparisons. For blood products with a concentration that was measured “too low to measure” according to the ELISA kit, the concentration was replaced with a 0.5 below the ELISA kit's documented “too low to measure” concentration (A2M: replaced with 0.38 pg/mL, A1AT: replaced with 12.39 pg/mL). Data was analyzed using R v. 4.5.3, including the RVAideMemoire package [18]. A p‐value < 0.05 was considered significant.

3. Results

A total of 24 horses were included in this study: Control‐group = 12 and NSAID‐group = 12. The Median ± SD [range] age for the control‐group was 9 ± 4.4 years [4 to 16 years] and included seven mares and five geldings, whereas the NSAID‐group was 14 ± 5.5 years [4 to 23 years] and included six mares and six geldings (Table 1). All data in this study were non‐normally distributed. One sample from the A2M analysis (n = 1 [NSAID‐liquid PRF]) had a “too low to measure” (TLTM) value and was replaced with 0.38 pg/mL. Visualization revealed that distributions of A2M concentration varied by product; while the distribution within plasma did appear to be approximately normally distributed, the distribution within solid PRF was highly left skewed, and within liquid PRF was highly right skewed. Seven samples from the A1AT analysis (n = 5 [NSAID‐plasma], n = 2 [Control‐plasma]) had a TLTM value and were replaced with 12.39 pg/mL. For blood products with TLTM values (A2M [liquid PRF] and A1AT [plasma]), their TLTM replacement values were used in the analysis.

Table 1 shows information for the 12 horses enrolled in the NSAID‐group including how many doses of flunixin meglumine they received in hospital before their blood samples were collected. Prior to using non‐parametric analyses to analyze the data, ICCs revealed poor to moderate relationships of antiprotease levels across the products, half of which were not statistically significant at the 0.05 level (Table 2) [19]. Table 3 shows the descriptive data and p‐values for the comparison between the control‐group and NSAID‐group for all three blood products (plasma, solid PRF, and liquid PRF), including A2M (liquid PRF) and A1AT (plasma) with and without TLTM values. No significant differences were found between the control‐group versus the NSAID‐group for any of the blood products (all p‐values > 0.05). Plasma had the highest concentration of A2M and was significantly higher than solid and liquid PRF (both p < 0.001). No significant difference was found between A2M in solid versus liquid PRF (p = 0.964) (Figure 1). Solid PRF had the highest amount of A1AT, and it was significantly higher when compared to liquid PRF (p < 0.001) but not compared to plasma (p = 0.071). A significant difference was also found between A1AT in plasma versus liquid PRF (p = 0.002), with liquid PRF having the lowest concentration of A1AT (Figure 1).

TABLE 2.

Intraclass correlation coefficients (ICC) comparing measures across blood products (plasma, solid platelet‐rich fibrin [PRF], liquid PRF) for the two antiproteases A2M (alpha‐2‐macroglobulin) and A1AT (alpha‐1‐antitrypsin).

Antiproteases blood products	ICC	p	Relationship
A2M Plasma versus solid PRF	0.064	0.380	Poor
A2M Plasma versus liquid PRF	0.519	0.004*	Moderate
A2M Solid PRF versus liquid PRF	0.124	0.278	Poor
A1AT Plasma versus solid PRF	0.639	< 0.001*	Moderate
A1AT Plasma versus liquid PRF	0.523	0.004*	Moderate
A1AT Solid PRF versus liquid PRF	0.155	0.230	Poor

Note: A black asterisk (*) indicates a significant p‐value and relationships are evaluated as poor < 0.5, moderate = 0.5 to 0.75, good = 0.75 to 0.9, and excellent > 0.9 [19].

TABLE 3.

The table shows the results for the two equine antiproteases alpha‐2‐macroglobulin (A2M) and alpha‐1‐antitrypsin (A1AT): Median and min (minimum)/max (maximum) concentrations for the three different blood products (plasma, solid PRF, liquid PRF) for the two different treatment groups (NSAID‐group, Control‐group).

A2M	Group	Median	Median* (number)	Min	Min*	Max	Max*	p*
Plasma	NSAID	17.31	NA	1.92	NA	32.56	NA	0.998
	Control	21.31	NA	7.42	NA	37.79	NA
Solid PRF	NSAID	15.03	NA	0.77	NA	18.91	NA	1.000
	Control	12.41	NA	3.26	NA	19.51	NA
Liquid PRF	NSAID	13.21	13.83 (11)	0.38	3.85	19.83	19.83	0.853 (0.333)
	Control	9.63	NA	1.57	NA	30.56	NA

A1AT	Group	Median	Median*	Min	Min*	Max	Max*	p*
Plasma	NSAID	82.11	141.8 (10)	12.39	26.73	667.69	707.23	0.773 (0.279)
	Control	95.75	180.34 (7)	12.39	42.4	707.23	667.69
Solid PRF	NSAID	157.35	NA	70.94	NA	1363.49	NA	0.483
	Control	230.15	NA	93.8	NA	9119.52	NA
Liquid PRF	NSAID	46.13	NA	24.78	NA	298.42	NA	0.599
	Control	69.08	NA	38.3	NA	597.95	NA

Note: Median*, min*, max*, and (p‐value)* = shows the result when the case number is adjusted from 12 to only include cases that did not measure “too low to measure.” The new case number for these cases can be found under Median* in the (). No significant differences were found between the two treatment groups (NSAID vs. Control) for any of the groups (all p‐values > 0.05).

Abbreviation: NA, not applicable.

FIGURE 1.

The median concentration differences for the two antiproteases alpha‐2‐macroglobulin (A2M) and alpha‐1‐antitrypsin (A1AT), between the three blood products: Plasma, solid platelet‐rich fibrin (PRF), and liquid PRF (black asterisk (*) = significance, ns = non‐significant, solid square = Control, solid triangle = NSAID, black square outline = Control “too low to measure,” black triangle outline = NSAID “too low to measure”).

4. Discussion

Systemic anti‐inflammatory medication, such as flunixin meglumine, is often a mainstay of treatment to reduce uveitis in cases of ulcerative keratitis in horses [1]. As such, many of the equine ophthalmology patients that serve as donors for autologous blood products may concurrently be receiving this systemic anti‐inflammatory medication at the time of collection. The present study found that a short course of systemic NSAIDs does not appear to alter antiprotease concentrations in autologous blood products to a meaningful degree. The median concentration of A2M and A1AT were comparable in both the control‐group and NSAID‐group for all three blood products. This suggests that autologous plasma or PRF from horses treated with a short course of anti‐inflammatory medication can still be used for treatment of keratomalacia without a presumed impact on antiprotease activity. Notably, the control groups showed the highest concentrations of both antiproteases, except for A2M in liquid PRF (Table 3).

Various antiprotease preparations exist in veterinary medicine for the treatment of keratomalacia. This includes the aforementioned autologous blood products (serum or plasma), as well as EDTA, acetylcysteine, tetracycline, and Ilomastat. Blood products provide natural antiproteases (A2M and A1AT) to decrease excessive activity of MMPs and NE, while the other antiproteases function by chelating calcium and zinc, cofactors necessary for proteolytic enzyme stabilization [7]. Studies have demonstrated that EDTA (0.2%), acetylcysteine (5%–10%), tetracycline (0.1%), and Ilomastat are potent MMP‐inhibitors but are not effective against serine proteases, such as NE [7, 20, 21]. Since horses with ulcerative keratitis have high concentrations of both MMPs and NE, combining various antiproteases to achieve a targeted and broad spectrum of activity if generally recommended [7]. Practically, this often includes an autologous blood product containing A2M and A1AT, combined with a metal‐ion chelator.

A2M is a broad‐spectrum, highly potent protease inhibitor that works by entrapping proteases, thereby preventing further corneal liquefaction [5, 7, 22, 23]. Due to its broad activity against proteases, including both MMPs and NE, and ability to trap two protease molecules per A2M molecule, high concentration preparations may be more beneficial against keratomalacia [22]. A1AT works mechanistically similar to A2M but may have a more limited spectrum [7]. A1AT is a serine protease inhibitor with known activity against NE but has more limited dose‐dependent MMP activity [7]. Thus, antiprotease selection containing high concentrations of both A2M and A1AT is theoretically superior in preventing progression of keratomalacia.

As a secondary objective, our study compared the concentrations of A2M and A1AT in plasma, solid PRF, and liquid PRF. Our study demonstrated that solid PRF has a high concentration of A1AT and, although not statistically significant compared to plasma, it does appear to be clinically high (Figure 1). Our research group has noticed that using solid PRF as a clot placed in the ventral fornix in melting corneal ulceration cases seems to decrease keratomalacia and firm up a melting ulcer, as described in a recently published case series [11]. This observation correlates with our study that shows that solid PRF has a clinically high concentration of A1AT and could therefore be as or more effective than plasma against melting ulcers. However, plasma was found to have significantly higher amounts of A2M compared to both solid and liquid PRF. Therefore, solid PRF is likely not an appropriate substitute for plasma but may be a beneficial adjunctive therapy due to its high A1AT concentration. In contrast, liquid PRF appears to be a relatively poor treatment modality for corneal melting when it comes to its antiprotease properties since it contains significantly lower concentrations of both A2M and A1AT compared to plasma. Future supporting studies may include further investigation of PRF release kinetics, stability, and direct evaluation of antiprotease efficacy on malacic corneas to better guide clinical recommendations.

Interestingly, there was substantial individual variability (range; min/max) in antiprotease concentrations among all three blood products in our study. This indicates that careful selection of horses as plasma donors for corneal ulceration patients is an important consideration. This large variation in antiprotease concentrations suggests that some horses may not be the most advantageous donors whereas others contain high concentrations of beneficial antiproteases. A study by Zimmerman and colleagues also found that protease concentrations were variable among individual horses but did not differ based on signalment (age, sex) [22]. A convenient, and fast bed‐side diagnostic (e.g., snap ELISA test) to evaluate antiprotease levels in blood products may be beneficial for the ophthalmology community to determine which horses would be good donors since there is no obvious predictability.

Although this study evaluated antiprotease concentrations in plasma, autologous serum has also been historically used for its antiprotease activity [24]. An in vitro study by Conway and colleagues found no difference in anti‐collagenase activity of plasma compared to serum, so these treatment modalities are often used interchangeably [25]. This is supported by other studies showing similar outcomes in decreasing protease activity [26, 27]. However, there may be other benefits of serum compared to plasma. A recent study in 2024 by Kubai and colleagues found that serum has greater antimicrobial effects [28]. Despite these findings, autologous plasma may be used in the clinical setting over serum due to convenience, since obtainment does not require coagulation of blood and larger quantities can be acquired [25]. Our study explicitly evaluated plasma, since this is what is utilized by our institution. However, for practitioners that use autologous serum, conclusions from this study should not be directly extrapolated to serum.

Various methods have been used to quantify antiprotease concentrations. Several older studies used protein electrophoresis or gelatin zymography [6, 29, 30]. However, these methods have a low specificity in which case they may overestimate concentrations [20]. Like our study, Zimmerman and Diehl used ELISA to measure A2M through highly specific antibody–antigen detection [22]. The ELISA kits in our present study also used analyte‐specific antibodies to detect A2M and A1AT. Recent studies have suggested that certain ELISAs may not contain an antibody that can bind A2M effectively, and thus mass spectrometry is proposed to be the better option for determining A2M concentrations due to its ability to detect concentrations below the level of ELISA detection [31, 32]. Our study had only one case in the NSAID group for liquid PRF in which A2M was below the level of ELISA detection. Meanwhile, there were seven cases in the A1AT analysis for plasma that were too low to measure. These findings support that no overarching conclusion can be drawn that ELISA is insufficient for measuring antiprotease concentrations in all circumstances. Rather, the utility of ELISA appears to be dependent on the methodology and test materials, which was also proposed by Ortved and colleagues [31]. In our study, the NSAID‐treated horses resulted in the most values below the level of ELISA detection for A1AT, particularly in plasma samples. Although this may represent a biological effect of flunixin meglumine, other underlying factors such as a limitation of the ELISA assay cannot be excluded. A future study could include using mass spectrometry as the analysis method to measure A2M and A1AT concentrations in plasma, solid PRF, and liquid PRF. This may provide more clarity on the effect of flunixin meglumine on A1AT concentrations in plasma as well as other preparations.

Our study was limited by sample size and an investigation with a larger number of horses may have increased our ability to detect significant differences. Additionally, the impact of systemic flunixin meglumine on antiprotease concentrations was only investigated after 3 days of systemic NSAIDs. Often, equine ophthalmic patients will be on systemic NSAIDs for even longer and the implications of a chronic course of flunixin meglumine on antiprotease concentrations in autologous plasma or PRF are still unknown. A future direction could include investigating if chronic anti‐inflammatory medication alters antiprotease concentrations in autologous plasma and PRF. Furthermore, other NSAIDs such as firocoxib and phenylbutazone were not included in the present study, but given their mechanism of action, similar outcomes are expected [33]. Additionally, this study compared antiprotease concentrations from independent populations, which may represent an inherent limitation of our study given the substantial individual variability identified in both our study as well as seen in Zimmerman and colleagues [22]. A longitudinal study measuring antiprotease concentrations pre‐ and post‐flunixin meglumine in individuals could reveal a more direct cause‐and‐effect relationship between flunixin meglumine and antiprotease concentrations and could be suggested as an additional direction for future studies.

In conclusion, a short course of systemic anti‐inflammatory medication in horses does not appear to alter antiprotease concentrations in autologous blood products such as plasma, solid PRF, and liquid PRF. However, substantial individual variability exists in antiprotease concentrations across blood products. Our study also suggests that although PRF should not be used as a replacement for plasma, solid PRF may be a beneficial adjunctive anti‐keratomalacia therapy to plasma given its high levels of A1AT. Future studies are warranted to support this hypothesis and guide clinical recommendations.

Author Contributions

Kaitlyn R. Scheunemann: conceptualization, investigation, writing – original draft, resources. Katrina E. V. Jones: conceptualization, investigation, writing – review and editing, methodology. Kimberly Young: investigation, writing – review and editing. Chongben Zhang: investigation, writing – review and editing, methodology. Cody McGrath: investigation, writing – review and editing, methodology. Kimberly R. Love: writing – review and editing, methodology, formal analysis. Michala de Linde Henriksen: conceptualization, investigation, funding acquisition, writing – original draft, methodology, validation, visualization, data curation, resources. All authors reviewed the results and approved the final version of the manuscript.

Ethics Statement

This study was approved by North Carolina State University's Institutional Animal Care and Use Committee (IACUC protocol #24‐348‐03). Animal owners or owners' representatives provided written informed consent for enrollment in the study, procedure(s) and therapy undertaken, and publication of data and images.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that supports the findings of this study are available on request from the corresponding author.