Feasibility of subantimicrobial-dose doxycycline for elbow tendinopathy

Abstract

Background

Elbow tendinopathy is a common condition among athletes and the general population and can result in significant disability and time loss from work. Despite traditional treatments, symptoms often persist for more than 6 months. Doxycycline has been suggested as having treatment potential for tendinopathy as a matrix metalloproteinase inhibitor. The objective of this study was to evaluate the feasibility of subantimicrobial dose (SD) doxycycline as a low-cost adjunctive treatment in the nonsurgical management of elbow tendinopathy.

Methods

Subjects received doxycycline 20 mg twice per day for 12 weeks in addition to standard of care, including a home exercise program and a counterforce brace. Any potential side effects were reported weekly, and the remaining pills were counted at 12 weeks. Patient-reported outcome measures were collected at baseline, 4, 8, and 12 weeks. Ultrasound, grip strength, and laboratory values were assessed at baseline and 12 weeks.

Results

Our sample consisted of 21 patients (average age 50.9 years, 57.1% male). The median weeks of elbow pain at baseline was 14 (interquartile range 15). SD doxycycline resulted in no significant side effects or changes to laboratory values. Median medication compliance was 96% by pill count. Patient-reported outcome measures and grip strength were recorded at multiple time points in the follow-up.

Conclusion

Patients with elbow tendinopathy tolerated SD doxycycline well and the majority were compliant with a twice-daily regimen for 12 weeks. Our cohort demonstrated improvements in symptoms and grip strength. Further research is needed to determine if SD doxycycline contributes to this improvement.

Tendinopathy is one of the most common conditions seen by sports medicine physicians, yet it remains one of the more challenging conditions to treat, especially in the case of lateral elbow tendinopathy.^20,32 Also referred to as tennis elbow or lateral epicondylitis, lateral elbow tendinopathy is relatively common within the general population, with an annual incidence ranging from 2 to 13 per 1000.^29,30 Nonsurgical management is initially recommended for lateral elbow tendinopathy, with treatments such as activity modification, nonsteroidal anti-inflammatories, physical therapy, bracing, and injections. The direct medical cost per patient associated with the treatment of lateral elbow tendinopathy is estimated to be $800 annually.²⁸ However, when combined with indirect costs, including time loss from work, the annual total cost of elbow tendinopathy in the United States has been estimated at over $22 billion.³⁷

Despite the various nonsurgical treatment options, nearly 1 in 5 patients with lateral elbow tendinopathy will require treatment for over 6 months and approximately 2% will eventually undergo surgery.^8,29 One study found that for patients with over 6 months of symptoms, the median duration of treatment was 844 days.²⁹ New and developing treatment options, such as platelet-rich plasma (PRP) injections and extracorporeal shock wave therapy (ESWT), have become increasingly utilized in recent years. Systematic reviews have mixed conclusions on the effectiveness of PRP injections for treating lateral elbow tendinopathy.^1,7 ESWT is similarly beneficial in single studies, but a Cochrane review showed no benefit.^3,14 The lack of insurance coverage for these newer treatments makes them cost-prohibitive for many patients. Therefore, a critical gap remains in the sports medicine clinician's ability to successfully treat elbow tendinopathy and allow patients to return to pain-free sports and occupational activities.

The pathophysiology of tendinopathy is multifactorial and contemporary studies demonstrate that this is associated with a failed healing response to repetitive mechanical loading, with gradual accumulation of matrix damage.⁹ Matrix metalloproteinases (MMPs), which are a family of endopeptidases, are known drivers of this pathophysiology by degrading the collagen matrix.^4,17,27 Doxycycline is a second-generation tetracycline that is used for both antibacterial and anti-inflammatory clinical indications. In addition to blocking bacterial protein translation, doxycycline has been shown to downregulate the production of proinflammatory cytokines, inhibit neutrophil chemotaxis, inhibit MMP activity, and promote connective tissue growth.^{13,15,21,34,35} Doxycycline is known to be a potent MMP inhibitor even at subantimicrobial doses.^10,25 Since patients with tendinopathy have elevated local MMP activity and a blunted matrix protein anabolic response to mechanical loading, doxycycline has a potential role in the treatment of chronic tendinopathy.⁶

Subantimicrobial dose (SD) doxycycline has been shown to have a positive effect in a variety of conditions, including acne, rosacea, periodontitis, and Graves' orbitopathy.^{10,19,23,25,33} These nonbacteriostatic indications require extended use of doxycycline, sometimes up to 2 years.^25,26 Studies of SD doxycycline also describe a low incidence of adverse reactions that is similar to placebo and without reports of photosensitivity.^{13,15,19,22,31,34} Furthermore, there does not appear to be an increased risk of antibiotic resistance or alteration in the fecal or vaginal microflora as a result of prolonged SD doxycycline use.^31,36

Doxycycline has been suggested as having a potential benefit in the treatment of tendinopathy as an MMP inhibitor.^9,12,18 In a rat model, oral doxycycline decreased MMP expression and improved tissue healing in the context of both an Achilles tendon tear and rotator cuff tear.^2,16,24 However, there are no published studies of the effect of doxycycline on tendinopathy in humans. The primary aim of this pilot study is to evaluate the feasibility of SD doxycycline as an adjunctive treatment for lateral and medial elbow tendinopathy. The purpose of this study is to evaluate the following objectives in order to plan for a future clinical trial:

• 1.Measure patient compliance in the treatment of lateral and medial elbow tendinopathy with SD doxycycline as well as tolerance to the medication.
• 2.Obtain data on patient-reported outcome measures (PROMs) and grip strength for patients with symptomatic lateral and medial elbow tendinopathy to allow planning of a future definitive efficacy trial.

Our central hypothesis for this pilot study is that patients with lateral and medial elbow tendinopathy will be receptive to treatment with SD doxycycline and be compliant with the prescribed treatment, with an acceptable side-effect profile.

Methods

Sample

This study was approved by the institutional review board of the Hospital for Special Surgery (HSS), and informed consent was obtained from all participants. The target population was otherwise healthy men and women, 18-60 years of age, with symptomatic lateral and medial elbow tendinopathy. Patients who had previously received noninvasive treatment (eg, nonsteroidal anti-inflammatories, physical therapy, and/or brace) were eligible. However, patients who had a prior elbow injection or surgery were excluded. Since doxycycline is contraindicated during pregnancy and while breastfeeding, we excluded pregnant women, women who were breastfeeding, and premenopausal women who were not using contraception.

All subjects were recruited through the practices of Drs. Toresdahl, DiFiori, Dines, and Rodeo, as well as the HSS Primary Sports Medicine Service. The sample size was calculated based on the estimated retention rate and compliance. A sample of 20 subjects would be able to estimate a retention rate of 80% with 15% absolute precision and 90% confidence. This would additionally allow for estimating an 80% compliance rate, which specifically refers to the number of pills taken (as measured by pill logs) divided by 168 (2 times daily × 7 days per week × 12 weeks).

Measures

Data were collected and managed using Research Electronic Data Capture (REDCap; Vanderbilt University, Nashville, TN, USA), which is a secure web platform for online databases and surveys supported by the Weill Cornell Medicine Clinical and Translational Science Center and subsidized by grant number UL1 TR0002384 from the National Center for Advancing Translational Sciences of the National Institutes of Health. Upon enrollment, subjects completed baseline questionnaires about demographics, hand dominance, symptom duration, activity, exercise, sport associated with the onset of symptoms, visual analog scale, prior treatments, relevant medical history, over-the-counter and prescription medication use, medication allergies, and for females, pregnancy or breastfeeding status. Subjects also completed the patient-rated tennis elbow evaluation (PRTEE) and quick Disabilities of the Arm, Shoulder and Hand score (qDASH) at baseline. The scoring ranges for these questionnaires are 0-100, where increased pain or disability is associated with a higher score. These instruments were selected given their reliability and validity in lateral and medial elbow tendinopathy with low respondent burden.¹¹ Visual analog scale, qDASH, and PRTEE were collected again at 4, 8, and 12 weeks, respectively.

Blood tests

Blood tests were performed at baseline and 12 weeks. Measured laboratory values included complete blood count, comprehensive metabolic panel, hemoglobin A1C, and C-reactive protein (CRP).

Grip strength

A Jamar hydraulic 90 kilogram (200 lb) hand dynamometer was used to assess grip strength in the affected extremity at baseline and at 12 weeks. The best of 3 efforts were recorded for both the affected and unaffected side.

Ultrasound

Tendinopathy grade was measured by grayscale ultrasound and with color Doppler at baseline and 12 weeks after enrollment. Scans were performed using either a GE Logiq e9 with a 10-15 MHz linear transducer (GE HealthCare, Chicago, IL, USA) or a Samsung RS 85 with a 12-18 MHz linear transducer (Samsung Medison Co., Seoul, South Korea). Elbow tendinopathy was assessed based on the presence of tendon hypoechogenicity, tear, increased thickness, cortical irregularity, intratendinous calcification, and enthesophyte as well as associated radial collateral ligament abnormality, as described by Chiavaras et al.⁵ All imaging for this phase of the study was performed by a board-certified radiologist (H.G.).

Study design

Doxycycline hyclate 20 mg twice daily was prescribed for 12 weeks. The dose of doxycycline was selected as the subantimicrobial dose that is available as a generic medication and commonly used for other indications where it has been shown to be well-tolerated and also decrease MMP and CRP levels.^25,34 To prevent malabsorption and esophagitis, subjects were advised to not take it at the same time as milk or dairy or calcium, iron, or magnesium supplements. Precautions regarding sun sensitivity were also reviewed, even though this side effect has not been reported with a subantimicrobial dose.³⁴ Compliance with the treatment protocol was measured using a self-report log, including pill logs, that were collected biweekly by the clinical research coordinator.

In addition to SD doxycycline, all subjects were provided standard-of-care nonsurgical treatment for elbow tendinopathy, including the following interventions:

• 1.Activity modification: Subjects were advised to minimize activities that exacerbate the pain.
• 2.Eccentric strengthening program: A home exercise program was provided focusing on eccentric strengthening of the wrist and forearm. Initial instructions were provided in person by the clinical research coordinator or certified athletic trainer at the time of study enrollment. Detailed handouts on the exercise program were provided along with a resistance bar of appropriate strength and resistance bands.
• 3.Counterforce brace: A counterforce brace was provided with instructions to use it as needed during the day with activities that cause moderate-to-severe pain (≥4 out of 10).
• 4.Pain control: For mild pain (1-3 out of 10), subjects were instructed to ice the elbow 20 minutes 2-3 times per day as needed. For moderate pain (4-7 out of 10), subjects were instructed to ice and take acetaminophen up to 1000 mg 3 times daily as needed as long as there was no contraindication to acetaminophen. For severe pain (8-10 out of 10), subjects were instructed to ice, take acetaminophen, and take ibuprofen up to 400 mg 3 times daily or naproxen up to 220 mg twice daily as long as there was no contraindication to acetaminophen, ibuprofen, or naproxen.

Statistical analysis

Descriptive statistics were used to assess patient characteristics and demographics. Follow-up scores for each PROM (ie, qDASH, PRTEE) and worst pain scores were measured at 4, 8, and 12 weeks and compared to baseline. Differences in scores between each follow-up point and baseline were assessed using Wilcoxon signed-rank tests. Differences in grip strength and laboratory between 12 weeks and baseline were also assessed using Wilcoxon signed-rank tests. Tests were 2-sided, and a significance level of 0.05 was used. Paired t-tests were used to assess differences in means for ultrasound tear dimension and tendon thickness at 12 weeks compared to baseline. Wilcoxon signed-rank tests were used to assess differences in the distributions of tendon echogenicity and hyperemia at 12 weeks compared to baseline. Finally, McNemar's test was used to assess differences in radial collateral ligament abnormality and enthesophyte presence at 12 weeks compared to baseline. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Frequencies and percentages of patient characteristics for the final sample are presented in Table I. A total of 21 patients were included in the study, of which 20 completed the 12-week protocol, corresponding to a retention rate of 96%. Our sample consisted of 57% male and 43% female patients. The majority of the sample was White (86%), followed by Black or African (10%), and other race (5%), which was representative of the Hospital's patient population. Of our sample of n = 21, 95% were non-Hispanic. The sample was primarily right-hand dominant (95%), and the dominant side was affected in 62% of patients. Among the 20 subjects that completed the study, the median medication compliance was 96% (162/168 by pill count, interquartile range: 15%), and the response rate for weekly logs was 98% (235/240).

Table I.

Baseline patient characteristics.

	n (%)
Total	21 (100%)
Age (mean ± SD)	50.86 ± 9.65
Sex
Male	12 (57.14%)
Female	9 (42.86%)
Race
American Indian or Alaska Native	0 (0.00%)
Asian	0 (0.00%)
Black or African	2 (9.52%)
Native Hawaiian or Pacific Islander	0 (0.00%)
White	18 (85.71%)
Other	1 (4.76%)
Ethnicity
Hispanic	1 (4.76%)
Non-Hispanic	20 (95.24%)
Affected side dominance
Dominant	13 (61.90%)
Nondominant	8 (38.10%)
Weeks of elbow pain at baseline (mean ± SD)	19.38 ± 14.44
Tendinopathy laterality
Medial	3 (14.29%)
Lateral	18 (85.71%)
Prior episodes of lateral elbow tendinopathy
No	16 (76.19%)
Yes	5 (23.81%)
Medical care received for current elbow pain
X-ray	12 (57.14%)
Ultrasound	8 (38.10%)
Magnetic resonance imaging	2 (9.52%)
Computed tomography scan	0 (0.00%)
Physical therapy or home exercise program	11 (52.38%)
Over-the-counter medication	8 (38.10%)
Prescription medication	1 (4.76%)
Massage therapy	2 (9.52%)
Acupuncture	2 (9.52%)
Other	1 (4.76%)

SD, standard deviation.

Supplementary Table S1 depicts potential side effects reported by patients in our study sample. Symptoms reported by patients potentially related to SD doxycycline treatment were fatigue (10%), headache (10%), and heartburn (10%). Table II illustrates responses to SD doxycycline recorded at 12-week follow up. The majority of patients either strongly agreed (76%) or agreed (12%) that they tolerated the SD doxycycline well. If found to be effective, 63% of patients indicated that they would take SD doxycycline again if they experienced another tendon injury. Further, 67% indicated that they would recommend SD doxycycline to a friend with a tendon injury.

Table II.

Patient responses to subantimicrobial dose (SD) doxycycline at 12-week follow-up.

	n	%
I tolerated SD doxycycline well
Strongly disagree	0	0.00
Disagree	0	0.00
Neutral	1	4.76
Agree	3	12.29
Strongly agree	16	76.19
Unknown	1	4.76
If found to be effective, I would take SD doxycycline again if I experienced another tendon injury
Strongly disagree	1	4.76
Disagree	0	0.00
Neutral	4	19.05
Agree	8	30.10
Strongly agree	7	33.33
Unknown	1	4.76
If found to be effective, I would recommend SD doxycycline to a friend with a tendon injury
Strongly disagree	1	4.76
Disagree	0	0.00
Neutral	5	23.81
Agree	7	33.33
Strongly agree	7	33.33
Unknown	1	4.76

Table III shows the median and interquartile range of worst pain, qDASH, and PRTEE scores at baseline and each follow-up point (ie, 4 weeks, 8 weeks, and 12 weeks) as well as P values associated with Wilcoxon signed-rank tests for score change. The median worst elbow pain at baseline was 72 compared to 21.5 at 12 weeks. The median qDASH at baseline was 25 compared to a median qDASH of 6.82 at 12 weeks. The median PRTEE at baseline was 29 compared to a median PRTEE of 7.25 at 12 weeks. Changes in worst pain, qDASH, and PRTEE were significant at all follow-up points. We also observed significant changes in grip strength (P = .0025) between baseline and 12 weeks. Conversely, we did not observe statistically significant changes in the majority of monitored laboratory values (ie, hemoglobin A1C, aspartate transferase, alanine transaminase, creatinine, blood urea nitrogen, hemoglobin, platelets, hematocrit, white blood cells, potassium, carbon dioxide, and CRP) between baseline and 12-week follow-up. However, statistically significant changes in sodium (P = .0003) and chloride (P = .0026) were noted, though these changes were not clinically significant. Supplementary Table S1 provides ultrasound measurements collected at baseline and 12 weeks. Of our total sample, 20 ultrasounds were performed at baseline, and 19 were performed at 12 weeks. The distribution of tendon echogenicity at baseline was not significantly different than the distribution at 12 weeks (P = .2500). At baseline, four tendons showed severe hyperemia compared to one at 12 weeks. The distribution of tendon hyperemia at baseline was significantly different than at 12 weeks (P = .0469). Additionally, four ultrasounds demonstrated cortical irregularity of the common extensor tendon at baseline compared to none at 12 weeks. The mean transaxial tear dimension was significantly greater at baseline than at 12 weeks (P = .0458); however, longitudinal tear dimension (P = .4611), transaxial tendon length (P = .2467), and transaxial tendon depth (P = .7989) were not significantly different at 12 weeks compared to baseline.

Table III.

Patient-reported outcome scores, grip strength, and laboratory values at baseline and follow-up (ie, 4 weeks, 8 weeks, and 12 weeks).

	Baseline (n = 21)	Four weeks (n = 21)	P value∗	Eight weeks (n = 19)	P value∗	Twelve weeks (n = 20)	P value∗
	Median (IQR)	Median (IQR)		Median (IQR)		Median (IQR)
Patient-reported outcomes
Worst elbow pain	72.00 (19.00)	50.00 (29.00)	.0002	27 .00 (12.00)	<.0001	21.50 (39.25)	<.0001
qDASH	25.00 (20.46)	20.45 (13.63)	.0198	9.09 (10.23)	.0004	6.82 (6.81)	.0002
PRTEE	29.00 (28.50)	22 (13.50)	.0048	8.00 (10.00)	.0001	7.25 (8.38)	.0002
Grip strength	12.71 (12.26)					16.34 (13.17)	.0025
Affected side	28.00 (27.00)	-	-	-	-	36.00 (29.00)	.0009
Laboratory values
HbA1c	5.30 (0.50)	-	-	-	-	5.20 (0.40)	.8351
AST	31.00 (9.00)	-	-	-	-	33.50 (11.50)	.2042
ALT	26.00 (18.00)	-	-	-	-	27.00 (18.50)	.3560
Creatinine	0.80 (0.10)	-	-	-	-	0.85 (0.10)	.1851
BUN	18.00 (6.00)	-	-	-	-	17.00 (6.50)	.8645
Hgb	13.90 (1.60)	-	-	-	-	13.95 (1.85)	.1355
PLT	226.00 (87.00)	-	-	-	-	231.00 (58.00)	.6873
HCT	41.40 (4.50)	-	-	-	-	40.75 (5.10)	.1297
WBC	5.70 (2.50)	-	-	-	-	6.05 (1.30)	.7214
Sodium	140.00 (3.00)	-	-	-	-	141.00 (3.00)	.0003
Potassium	4.50 (0.40)	-	-	-	-	4.50 (0.50)	.8087
Chloride	103.00 (3.00)	-	-	-	-	105.00 (3.50)	.0026
CO2	28.00 (3.00)	-	-	-	-	27.00 (3.50)	.6864
CRP	0.50 (0.20)	-	-	-	-	0.50 (0.20)	.9999

IQR, interquartile range; qDASH, quick disabilities of the arm, shoulder, and hand; PRTEE, patient-rated tennis elbow evaluation; HbA1c, hemoglobin A1C; AST, aspartate transferase; ALT, alanine transaminase; BUN, blood urea nitrogen; Hgb, hemoglobin; PLT, platelets; HCT, hematocrit; WBC, white blood cells; CO2, carbon dioxide; CRP, C-reactive protein.

Bold values indicate statistical significance (P-value < .05).

^∗Wilcoxon signed-rank test.

Discussion

This pilot study demonstrated that among patients with lateral and medial elbow tendinopathy, SD doxycycline is well-tolerated with a low prevalence of side effects. Monitored laboratory values of blood glucose, liver enzymes, and kidney function were not significantly changed. Most patients reported being willing to take this medication for a future tendon injury if the treatment was later proven effective, and most patients were compliant with the study protocol.

In addition to obtaining preliminary data on medication compliance and tolerance, relevant outcome measures were collected, including standard PROMs, grip strength, and ultrasound, which will allow for the design of a future clinical trial. Therefore, further controlled investigation is indicated. The natural history of elbow tendinopathy is that the condition generally improves with time. In addition, our study assessed SD doxycycline as an adjunctive treatment to standard-of-care nonsurgical treatments of elbow tendinopathy, including counterforce bracing and exercise programs. Thus, though it is plausible that the SD doxycycline had a positive effect given its known role as an MMP inhibitor, we are unable to draw causal conclusions about this treatment from this study as it is purely a pilot study with no control group, and any or all the interventions utilized (ie, counterforce brace, home exercise program, SD doxycycline, pain management strategies) could be contributing to the reduction in symptoms observed.

As a generic medication, the cost of SD doxycycline is affordable, thereby making this treatment more accessible than PRP injections or ESWT. Prescribing SD doxycycline does not require specialized procedural training as is necessary for PRP injections and ESWT, which also makes it more accessible as it could be prescribed by primary care providers in rural or underserved communities. Thus, SD doxycycline could be a powerful and accessible treatment for many individuals with lateral and medial elbow tendinopathy.

This prospective pilot study justifies further investigation of this treatment within this and similar patient populations. Given the feasibility of SD doxycycline shown by this study, a phase II trial can be performed to evaluate the efficacy of SD doxycycline for lateral and medial elbow tendinopathy. Future studies could also investigate the effect of SD doxycycline on other common types of tendinopathy. Alternative doses and delivery methods could also be explored, such as topical application via transdermal patch or gel. In future research, it may also be possible to track MMP activity, as it is associated with the SD doxycycline treatment. Finally, future research could also use ultrasound to evaluate the effect of SD doxycycline treatment on changes in tendon microstructure and composition.

This study is impacted by a few important limitations. First, our cohort consisted of patients of the clinician investigators and HSS Primary Sports Medicine Service, which may limit the generalizability of our findings. Importantly, our sample consisted of a higher proportion of non-Hispanic and White patients compared to the general population, indicating that our cohort under-represents minority populations. Additionally, as part of our protocol, we provided a home exercise program and counterforce brace for all study participants, so we are unable to determine how much of the improvements in pain observed are attributed to the SD doxycycline. Results may be confounded by concurrent use of home exercise programs, counterforce bracing, and anti-inflammatory medications. As this study was a prospective pilot study and not a randomized controlled trial, our study is unable to draw causal conclusions about the SD doxycycline treatment.

Conclusion

This study demonstrates that treatment of elbow tendinopathy with SD doxycycline is practical, feasible, and safe. The majority of patients were compliant with this adjunctive treatment protocol, and patients tolerated the medication well. The results of this study can guide the design of a prospective randomized controlled trial to evaluate the effectiveness of SD doxycycline for lateral and medial elbow tendinopathy. Our cohort demonstrated improvements in symptoms and grip strength. Further research is needed to determine if SD doxycycline contributes to this improvement.

Acknowledgment

The authors thank Hospital for Special Surgery for support through the HSS Surgeon-in-Chief Grant and the faculty of the HSS Primary Sports Medicine Service for assistance in recruiting subjects, specifically Drs. Marci Goolsby, Ryan Lingor, Daphne Scott, David Wang, and Warren Young.

Disclaimers:

Funding: Internal funding was used to conduct this study.

Conflicts of interest: BT: Research Support from the American Medical Society for Sports Medicine Collaborative Research Network Grant. Associate Editor for BJSM. Team Physician for the United States Biathlon Association. These disclosures are not applicable to the current work. JD: Royalties from Arthrex and ConMed Linvatec. Consultant/Speakers' Bureau for Arthrex. Ownership interest in ViewFi Health. These disclosures are not applicable to the current work. AM: Unpaid Executive Board membership for the Musculoskeletal Infection Society. This disclosure is not applicable to the current work. MO: Ownership interest in Jannu Therapeutics. This disclosure is not applicable to the current work. SR: Research Support from the Virginia Toulmin Foundation, Orthopaedic Research and Education Foundation, Arthritis Foundation, Angiocrine Biosciences, Inc., Weill Cornell Clinical & Translational Science Center, and National Institutes of Health. Consultant for Novartis Pharmaceuticals and Advance Medical. Ownership interest in Jannu Therapeutics, Overture Medical, and ChitogenX. These disclosures are not applicable to the current work. The authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.

Data sharing statement

The data utilized in this manuscript are not publicly available, but deidentified data are available upon request from the authors. Please contact Brett G. Toresdahl (brett.toresdahl@hsc.utah.edu) for more information.