When Lateral Epicondylitis Is Not Lateral Epicondylitis: Analysis of the Risk Factors for the Misdiagnosis of Lateral Elbow Pain

Methods

From May 2019 to December 2023, a total of 216 patients were evaluated at our clinic with a previous diagnosis of lateral epicondylitis. These patients had experienced partial or complete failure of nonoperative treatment for lateral epicondylitis and were referred to our outpatient clinic for a secondary consultation by their general practitioner or physical therapist. Inclusion criteria were as follows: a previous diagnosis of lateral epicondylitis or tennis elbow; a minimum of 6 months of symptoms; ultrasound or magnetic resonance imaging (MRI) suggesting degenerative changes in the common extensor tendon, indicative of lateral epicondylitis; and partial or complete ineffectiveness of at least 3 months of nonoperative treatment. Patients with an incomplete set of data available for analysis were excluded.

The duration of symptoms before the examination was recorded. For those patients who reported a recurrence of lateral elbow pain after the complete resolution of pain, the total number of months in which they suffered from lateral pain was reported.

The following epidemiological data and medical history details were collected: age, sex, occupational type (sedentary vs heavy work), prolonged use of a computer mouse on the same side as elbow pain, pain in the dominant arm, history of elbow trauma, engagement in sports considered at risk for lateral epicondylitis (such as tennis, golf, paddle tennis ¹⁹ ), history of elbow swelling or a diagnosis of inflammatory arthropathy, history of temporary or persistent limited range of motion (ROM), history of temporary or persistent mechanical symptoms (clicking/snapping/crepitus/locking sensation), and number of cortisone injections administered.

Conducted by the senior author (D.B.), the physical examination comprised an array of assessments: an evaluation of ROM, ² palpation of the lateral elbow to delineate the major pain location as typical or atypical (Figures 1 and 2), the Cozen test, ¹⁷ a test for posterolateral plica,^1,11 the identification of any mechanical symptoms during motion (clicking/snapping/crepitus/locking sensation), pain upon palpation at the level of the PIN located 3 to 4 cm distal to the lateral epicondyle (Figure 2B), an assessment of weakness during repetitive (at least 3 repetitions) resisted active supination, and the posterolateral drawer test (clearly positive, slightly positive but clearly different from the contralateral side, negative). To conduct the Cozen test, the affected elbow was kept fully straight with the forearm in pronation. The examiner secured the elbow by placing his or her thumb on the lateral epicondyle. The patient was then directed to clench his or her fist tightly and extend the wrist against resistance ¹⁷ ; the presence of pain over the lateral aspect of the elbow joint indicated a positive result.

OPEN IN VIEWER

Figure 1.

Major pain upon palpation at the lateral epicondyle or slightly anterior to it was considered a symptom of typical lateral epicondylitis.

OPEN IN VIEWER

Figure 2.

Atypical lateral elbow pain. The red circles show the atypical lateral elbow pain locations. In case of concomitant pain in different areas, the point of greatest pain was considered. (A) Pain suggestive of a radial nerve origin for lateral elbow pain. (B) Pain suggestive of the posterior interosseous nerve as the origin for lateral elbow pain. (C) Pain suggestive of a triceps origin for lateral elbow pain. (D) Pain suggestive of a posterolateral plica or lateral collateral ligament lesion as the origin for lateral elbow pain.

During the posterolateral drawer test, some patients exhibited a spontaneous and involuntary succession of contraction and relaxation of the anconeus muscle, which was visible beneath the skin. This phenomenon was interpreted as a fasciculation of the anconeus. A fasciculation of the anconeus during the posterolateral drawer test was recorded as a separate phenomenon but not as a positive test finding. Increasing pain in elbow flexion and/or extension was also recorded.

On the basis of medical history and clinical findings, patients were preliminarily classified into 2 groups: those with typical lateral epicondylitis and those with preliminary atypical lateral epicondylitis. Patients in the typical lateral epicondylitis group reported no history of trauma, no swelling, and no history of temporary or persistent limited ROM. On physical examination, patients in the typical lateral epicondylitis group exhibited normal ROM, pain upon palpation at the anterolateral aspect of the lateral epicondyle (Figure 1), a positive Cozen test finding, a stable elbow, an absence of strength loss during resisted supination, an absence of mechanical symptoms, no swelling, and no PIN pain. These patients were confirmed as having lateral epicondylitis and were advised to continue nonoperative treatment or were scheduled for a surgical intervention.

Patients were classified as the preliminary atypical lateral epicondylitis group if they presented at least one of the following attributes in their medical history: elbow trauma, limited ROM or mechanical symptoms, atypical pain (Figure 2), history of elbow joint swelling, or diagnosis of inflammatory arthropathy. On physical examination, patients in the atypical lateral epicondylitis group showed at least one of the following features: atypical pain including PIN pain, a negative Cozen test finding, signs of instability during the posterolateral drawer test, weakness during resisted supination, mechanical symptoms or restricted ROM, and a positive or dubious test finding for posterolateral plica. Patients in the atypical group underwent MRI, ¹⁴ computed tomography, and/or electromyography (EMG) according to suspected underlying abnormalities. In some patients, the final diagnosis was confirmed after intraoperative analysis of tissue samples (Appendix Table A1 [available in the online version of this article]).

After reviewing second-level imaging and, when indicated, performing intraoperative analysis of tissue samples, patients were ultimately categorized as having typical lateral epicondylitis, atypical lateral epicondylitis, or no lateral epicondylitis (misdiagnosis) (Figure 3). The last category included patients without lateral epicondylitis as the primary source of elbow pain. Also included in the misdiagnosis category were patients with concomitant lateral epicondylitis that was not thought to be responsible for the main symptoms at the time of patient enrollment in the study.

OPEN IN VIEWER

Figure 3.

Study flowchart.

Results

Of the 216 patients initially included, 25 were excluded for having symptoms for <6 months, leaving 191 patients for analysis. Among these 191 patients, 52 were initially classified as having preliminary atypical lateral epicondylitis (Figure 3). There were 2 patients in the preliminary atypical lateral epicondylitis group who were excluded because they did not follow through with requested second-level imaging for a final diagnosis. Of the remaining 189 patients, 21 (11%) were found to have lateral elbow pain not due to lateral epicondylitis (misdiagnosis group). Figure 3 shows a visual representation of the study’s flowchart.

Of the 21 patients in the misdiagnosis group, 6 (29%) were diagnosed with posterolateral elbow instability, 3 (14%) with PIN compression/irritation, 3 (14%) with inflammatory osteoarthritis (1 rheumatoid arthritis, 2 psoriasis), 2 (10%) with osteochondritis dissecans of the elbow, 2 (10%) with posterolateral plica, and 2 (10%) with primary osteoarthritis. Other final diagnoses were the presence of loose bodies in the elbow joint (n = 1), pigmented villonodular synovitis (n = 1), and a partial triceps lesion (n = 1). One of the patients with posterolateral instability also had some concomitant loose bodies. The details of the 21 misdiagnosed patients are reported in Appendix Table A1.

Table 1 presents the results of univariate analysis comparing the misdiagnosis group and the epicondylitis group. After adjusting for potential confounding factors using logistic regression analysis, the following factors were finally identified as being significantly associated with a misdiagnosis: history of trauma (OR, 17.85 [95% confidence interval (CI), 4.43-60.78]; P = .0027), history of ROM limitation and/or mechanical symptoms (OR, 16.68 [95% CI, 13.80-280.97]; P = .0278), history of swelling (OR, 14.32 [95% CI, 8.63-132.24]; P = .0032), and number of corticosteroid injections (OR, 2.00 [95% CI, 0.81-2.03]; P = .0007).

OPEN IN VIEWER

Table 1

Patient Characteristics ^a

	Epicondylitis Group (n = 168)	Misdiagnosis Group (n = 21)	t Value (95% CI)	P Value
Age, y	48.20 ± 9.23	43.95 ± 15.93	−1.44 (–8.20 to 1.27)	.15
Symptom duration, mo	13.96 ± 5.44	13.57 ± 5.65	−0.31 (–2.88 to 2.11)	.76
No. of corticosteroid injections	1.63 ± 1.23	3.05 ± 2.04	1.87 (0.81 to 2.03)	<.001
	Epicondylitis Group (n = 168)	Misdiagnosis Group (n = 21)	χ2 Value	P Value
Sex			0.70	.40
Female	43	33
Male	57	67
Heavy or repetitive work	66	29	11.13	<.001
Routine mouse use	46	43	0.10	.76
Engagement in at-risk sports	26	48	4.20	.04
Dominant arm affected	53	71	1.88	.11
History of trauma	2	29	25.55	<.001
History of ROM limitation or mechanical symptoms	1	29	34.52	<.001
History of swelling	2	38	44.90	<.001

a

Data are presented as mean ± SD or percentages. ROM, range of motion.

Post hoc analysis was performed for the variable age, which showed a bimodal distribution (Figure 4). Patients aged ≤30 years showed higher rates of lateral pain not due to lateral epicondylitis (OR, 66.90; P < .001). Patients aged >50 years did not show any significantly increased risk of a misdiagnosis (OR, 1.05; P = .90).

OPEN IN VIEWER

Figure 4.

Bimodal distribution of the rate of misdiagnoses at final diagnosis ranked by age group. All patients aged ≤20 years (3/189) did not have a final diagnosis of lateral epicondylitis. They were diagnosed with osteochondritis dissecans (n = 1), posterolateral plica (n = 1), and a triceps lesion (n = 1).

The analysis of clinical tests is presented in Table 2. The most sensitive test able to unveil which patients were not affected by lateral epicondylitis was the assessment of pain location. Atypical pain highly predicted a misdiagnosis, with a sensitivity of 90.5% (95% CI, 70%-99%). The absence of a positive Cozen test finding showed a sensitivity of 61.9% (95% CI, 38%-82%). ¹⁷ All the other tests displayed a high specificity but a low sensitivity or viceversa.

OPEN IN VIEWER

Table 2

Analysis of Clinical Tests ^a

	Sensitivity	Specificity	Positive Predictive Value	Negative Predictive Value	Accuracy
Atypical palpatory pain	90	99	95	99	99
Pain increase during elbow extension	57	3	7	36	9
Pain increase during elbow flexion	43	97	64	93	91
ROM limitation	43	99	82	93	93
PIN pain	29	88	24	91	81
Mechanical symptoms during ROM	24	99	83	91	91
Positive drawer test finding	24	99	83	91	91
Negative Cozen test finding	62	98	76	95	94

a

Data are presented as percentages. PIN, posterior interosseous nerve; ROM, range of motion.

Fasciculations of the anconeus during the posterolateral drawer test were observed in 19 patients (10%). The rate of fasciculations was higher in the atypical lateral epicondylitis group (24% [7/29]) and in the misdiagnosis group (29% [6/21]) compared with the rate observed in the typical lateral epicondylitis group (4% [6/139]) (P < .001).

Discussion

This study investigated, for the first time, the rate of misdiagnoses of lateral elbow pain at a referral shoulder and elbow center, involving 189 patients with suspected lateral epicondylitis who had at least 6 months of persistent symptoms. The 6-month limit was used to reduce the study population to those patients who were more likely to have an incorrect diagnosis. The prospective design of the study necessitated this limit to focus the researchers’ efforts on patients at a higher risk of a misdiagnosis. In this study population, there was an 11% rate of misdiagnoses. The most common differential diagnoses of lateral elbow pain were posterolateral instability, PIN compression/irritation, and inflammatory arthritis.

Among the 6 patients with posterolateral elbow instability, only 2 reported elbow trauma. In the remaining nontraumatic painful elbows, we found a median of 5 corticosteroid injections, suggesting potential iatrogenic damage due to repetitive injections in the lateral collateral ligament. This corroborates what has been reported by Chanlalit and Dilokhuttakarn. ⁵ The distinction in cases of lateral elbow pain is important because different diagnoses require distinct nonoperative or operative treatment. For example, lateral elbow pain associated with posterolateral elbow instability can be worsened by surgical treatment such as arthroscopic or open release of the extensor tendon.

PIN compression has been widely described as a potential cause of lateral elbow pain, ¹⁰ but its pathogenesis and diagnosis are unclear. ¹² In our cohort, clinical PIN compression was present in 3 patients. All of the 3 patients had clear weak resistance against supination and pain evoked by compression at about 3 to 4 cm distal to the epicondyle. In 2 cases, pain was also elicited when applying proximal compression near the radial head (Figure 2A). EMG findings were abnormal and indicative for PIN damage in 1 patient and unclear in 1 patient. One patient refused to undergo EMG.

Pain at the level of the PIN during the examiner’s compression, but not supination weakness, was much more common. Overall, 20 patients had PIN pain concomitant with lateral epicondylitis symptoms and were classified as having atypical lateral epicondylitis. The reason for this distal pain remains unclear and could be related to damage of unmyelinated nerve fibers due to extrinsic, but not detectable, compression or due to central sensitization of chronic pain.^6,9 The mean duration of lateral elbow pain in our series was more than 1 year, suggesting that central sensitization of chronic pain could play a certain role in the perpetuation of pain, despite treatment. Bonczar et al ³ reported that PIN compression can frequently be a coexisting cause of lateral elbow pain, and our data seem to confirm this association.

Inflammatory arthropathy of the elbow due to psoriasis and/or rheumatoid arthritis was observed in 3 patients. Although rheumatoid arthritis is typically a disease of the small joints, it can affect the elbow in 20% to 65% of cases. ⁸ None of the 3 patients with inflammatory arthropathy reported in our series had a previous diagnosis of any inflammatory disease. This made the diagnosis challenging and required a high degree of clinical suspicion and close cooperation with radiologists and rheumatologists. In their medical history, all the patients with rheumatoid arthritis reported episodes of joint effusion, and this was key in suspecting an inflammatory disease. On physical examination, they all showed some minor loss of extension as well as clear joint effusion on MRI.

This study strongly suggests that careful medical history taking is mandatory to avoid misdiagnoses in cases of longstanding lateral elbow pain. Patients with a history of elbow trauma had an OR for a misdiagnosis of 17.85, strongly suggesting that in case of previous elbow trauma, the evaluating doctor should have a high degree of suspicion for different and misleading causes of lateral elbow pain, such as posterolateral elbow instability, posttraumatic osteoarthritis, and posterolateral plica, especially in young patients.

Age is, in fact, another important variable that must be considered as a red flag. In our series, all the patients aged ≤20 years, and the majority of the patients aged between 21 and 30 years, did not have a final diagnosis of lateral epicondylitis. The rate of misdiagnoses in young patients was so high in our series as to lead us to conclude that lateral elbow pain in young patients should be considered non–lateral epicondylitis pain until proven otherwise.

A meticulous examination of the painful elbow is key to avoid a misdiagnosis. Patients with atypical pain should be highly suspected for a misdiagnosis. In this study, the presence of atypical pain had a sensitivity for a misdiagnosis of 90.5%. A negative Cozen test finding was moderate, suggestive of a misdiagnosis, with a sensitivity of 61.9%. Considering their sensitivity, the assessment of atypical pain and the Cozen test (negative finding) should be used as screening tests.

All the other tests showed a low sensitivity and a high specificity, except for pain increase during extension, which exhibited very low specificity. This was expected, considering that they are designed to make a diagnosis of specific elbow abnormalities such as posterolateral elbow instability or posterolateral plica. In simple terms, a positive posterolateral drawer test finding is highly predictive of a true lesion of the lateral collateral ligament, but a negative finding does not prove that there are no other causes of a lateral elbow misdiagnoses.

The presence of fasciculations of the anconeus during the posterolateral drawer test deserves further discussion. This phenomenon, to the best of our knowledge, has never been reported in the literature. The pathological background of a fasciculation is not known and requires further investigation. One of the hypotheses argues that overstretching of the anconeus’ fibers and its neurovascular bundle, which occurs during the drawer test in elbows affected by underlying subtle instability, determines the onset of fasciculations. This hypothesis seems to be supported by the high rate of fasciculations detected among the patients with posterolateral instability (5/6 [83%]). The observation of fasciculations at the level of the anconeus, being a phenomenon that has never been reported before, must be evaluated with caution. Similar to all clinical tests, its assessment is based on the interpretation of a phenomenon or symptom and is therefore subject to errors. Future studies are needed to further investigate this phenomenon.

Benign fasciculations have been reported after the use of corticosteroids. ¹³ Multiple injections at the level of the epicondyle could be primarily or jointly responsible for such phenomena.

This study has some limitations. The diagnosis of lateral epicondylitis was not confirmed surgically and was primarily based on clinical presentation, a physical examination, and imaging. The rate of misdiagnoses reported in this study should not be generalized because this rate was affected by several variables including the prevalence of specific elbow abnormalities, which could be different in different countries.

Moreover, it is important to note that a misdiagnosis does not necessarily equate to an error. An example could be the diagnosis of posterolateral elbow instability; some of the patients with a previous correct diagnosis of lateral epicondylitis subsequently may have developed posterolateral instability due to repeated intra-articular steroid injections. Despite these limitations, this study described, for the first time, the rate of misdiagnoses of lateral elbow pain, making the effort to define which patients are at risk for a misdiagnosis and to provide the physician with instruments to make a correct and prompt diagnosis.

Conclusion

The rate of misdiagnoses of lateral elbow pain in our population was 11%. A high degree of suspicion is mandatory in patients aged ≤30 years and in cases of trauma, swelling, ROM limitation and/or mechanical symptoms, and repetitive corticosteroid injections. An assessment of atypical pain location is a valuable screening test to select those patients at risk for a misdiagnosis.

Supplemental Material