Studied combinations

Spurling's test has been studied as one of four items in a cluster for cervical radiculopathy (Wainner 2003): upper limb tension test A positive, neck rotation of less than 60° towards the painful side, symptoms eased by neck distraction, and Spurling's test positive. In the original study of 82 patients (reference standard: nerve conduction and EMG testing), three of the four positive gave an LR+ of 6.1 and all four positive gave an LR+ of 30.3, with very wide confidence intervals (figures from secondary sources; original full text not seen). An independent study in 2026 of 85 patients (reference standard: MRI plus specialist diagnosis) found an LR+ of 7.7 (95% CI 3.2–20.7) with three of the four positive; all four positive picked up 18% of those with radiculopathy, with no false positives. These figures come from the tests as performed and scored in those studies.

Sources: Wainner RS et al. Spine 2003;28(1):52–62. Grondin F et al. Braz J Phys Ther 2026;30(3):101581.

Hoffmann's sign has been studied as one of five findings in a cluster for cervical myelopathy (Cook 2010): gait deviation, Hoffmann's sign, inverted supinator sign, Babinski sign, and age over 45. In 249 patients, three or more of the five positive gave an LR+ of 30.9 (95% CI 5.5–181.8). The five items were chosen from 13 findings in the same patients, so the figures come from the study that built the cluster; no independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Cook C et al. J Man Manip Ther 2010;18(4):175–180.

The Hawkins–Kennedy test has been studied as one of three tests in a cluster for subacromial impingement (Park 2005): Hawkins–Kennedy, painful arc, and the infraspinatus (resisted external rotation) test. In 552 patients having shoulder arthroscopy (reference standard: arthroscopy; 65% had some degree of impingement, from bursitis to full-thickness tear), all three positive gave an LR+ of 10.56, two of the three an LR+ of 5.03, and all three negative an LR− of 0.17; no confidence intervals were seen (figures from secondary sources; original full text not seen). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Park HB et al. J Bone Joint Surg Am 2005;87:1446–55.

Neer's sign has been studied as one of five tests in a cluster for subacromial impingement (Michener 2009): Neer's, Hawkins–Kennedy, painful arc, empty can and resisted external rotation. In 55 patients from a surgeon's shoulder clinic (reference standard: diagnosis at surgery; 16 had impingement), three or more of the five positive gave a sensitivity of 75% (95% CI 54–96%) and a specificity of 74% (61–88%), an LR+ of 2.93 (1.60–5.36) and an LR− of 0.34 (0.14–0.80) (cut-off from the abstract; figures from secondary sources; original full text not seen). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Michener LA et al. Arch Phys Med Rehabil 2009;90:1898–903.

The cross-body adduction test has been studied as one of four tests for acromioclavicular joint pain in primary care (Cadogan 2013): cross-body adduction, O'Brien's active compression test, the Hawkins–Kennedy test and AC joint tenderness. In 153 patients analysed (208 enrolled; reference standard: at least 80% pain relief after an AC joint injection), all four positive gave a sensitivity of 5% and a specificity of 99% (LR+ 5.70, 95% CI 0.60–52.63), and none of the four positive gave an LR− of 0.65 (0.11–3.54). Overall, the number of positive tests did not separate patients with and without AC joint pain (area under the curve 0.51). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Cadogan A et al. BMC Musculoskelet Disord 2013;14:156 (Table 4).

The compression test has been studied as one of six provocation tests for sacroiliac joint pain (Laslett 2005: distraction, thigh thrust, compression, sacral thrust, and Gaenslen's test on each side), against pain relief from an anaesthetic injection into the joint. In 48 patients analysed (62 recruited), three or more of the six positive gave a sensitivity of 94% and a specificity of 78% (published LR+ 4.29, 95% CI 2.34–8.58); the LR− works out at about 0.08. In the same study, two or more positive of four tests (distraction, thigh thrust, compression, sacral thrust) gave a sensitivity of 88% and a specificity of 78% (LR+ 4.00, 95% CI 2.13–8.08; LR− 0.16, 95% CI 0.04–0.47). A 2021 review pooling five studies of provocation-test clusters found an LR+ of 2.13 (95% CI 1.2–3.9) and an LR− of 0.33 (0.11–0.72) (abstract source only). These figures come from the tests as performed and scored in that study.

Sources: Laslett M et al. Man Ther 2005;10(3):207–218 (Tables 4 and 6). Saueressig T et al. J Orthop Sports Phys Ther 2021;51(9):422–431.

Reduced passive hip adduction has been studied as one of seven findings for severe hip osteoarthritis on X-ray (Bierma-Zeinstra 2002): age over 60, inguinal ligament tenderness, reduced external rotation, reduced internal rotation, reduced adduction, bony restriction of passive movement, and abductor weakness. In 220 primary-care patients aged 50 or over with hip pain (reference standard: X-ray, joint space 1.5 mm or less; 11.4% had severe OA), five or more present gave a sensitivity of 72% and a specificity of 98% (abstract). A 2019 review recalculated likelihood ratios from the study's data: five or more, LR 35 (95% CI 13–95); three or fewer, LR 0.05 (0.01–0.32) (original full text not seen). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Bierma-Zeinstra SM et al. J Rheumatol 2002;29:1713–8; Metcalfe D et al. JAMA 2019;322:2323–33.

The scour test (with adduction) has been studied as one of five findings in a rule for hip osteoarthritis on X-ray (Sutlive 2008) — a different target from the intra-articular hip pain figures on this page: hip pain on squatting, a positive scour test with adduction, lateral hip pain on active hip flexion, pain on active hip extension, and passive internal rotation of 25° or less. In 72 patients analysed (78 recruited; reference standard: standing X-ray, Kellgren–Lawrence grade 2 or more; 29% had OA), four or more of the five present gave an LR+ of 24.3 (95% CI 4.4–142.1), and three or more an LR+ of 5.2 (four-or-more figures from the abstract; item list, number recruited and three-or-more figure from secondary sources; original full text not seen). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Sutlive TG et al. J Orthop Sports Phys Ther 2008;38:542–50.

The active piriformis test has been studied in combination with the seated piriformis stretch test for sciatic nerve entrapment (Martin 2014). In 33 patients (23 with entrapment found at endoscopy and 10 without; reference standard: endoscopic assessment, reviewed retrospectively), the two tests combined gave a sensitivity of 91%, a specificity of 80%, an LR+ of 4.57 and an LR− of 0.11 (abstract source only; the abstract does not state whether one or both tests had to be positive, and no confidence intervals were seen). No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Martin HD et al. Knee Surg Sports Traumatol Arthrosc 2014;22:882–8.

The valgus stress test at 30° has been studied in a combination for medial collateral ligament (MCL) lesions in general practice (Kastelein 2008): history findings (trauma by an external force to the leg, rotational trauma) plus pain and laxity on valgus stress at 30°. In 134 patients aged 18–65 seen within 5 weeks of a knee injury (reference standard: MRI; 35 had an MCL lesion), adding pain and laxity on the test to the history gave an LR+ of 6.4 (abstract source only; no confidence interval seen). The items were selected from the same patients' data, and no independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: Kastelein M et al. Am J Med 2008;121:982–8.

The anterior drawer test has been studied as part of a delayed physical examination for lateral ankle ligament rupture (van Dijk 1996): bruising, site of tenderness and the anterior drawer test, assessed about 5 days after an inversion injury. In 160 patients (reference standard: findings at surgery in 135, and at least 6 months' follow-up in 25), this examination by an experienced orthopaedic surgeon gave a sensitivity of 96% and a specificity of 84%, with no confidence intervals published; the LR+ works out at about 6.0 and the LR− at about 0.05. Tenderness over the anterior talofibular ligament, bruising and a positive drawer together had a sensitivity of 96%; no specificity was published for that combination. No independent validation study was found. These figures come from the tests as performed and scored in that study.

Sources: van Dijk CN et al. Acta Orthop Scand 1996;67:566–70.

Snuffbox tenderness has been studied as one of three signs combined for scaphoid fracture: snuffbox tenderness, scaphoid tubercle tenderness and pain on thumb compression. In 215 patients with a suspected scaphoid fracture (Parvizi 1998), the three signs combined within 24 hours of injury gave a sensitivity of 100% and a specificity of 74% (abstract source only); the LR+ works out at about 3.8. In 154 patients with a normal first X-ray (Bergh 2014, MRI reference), a weighted score of the same three signs (4 or more of 6 points) gave a sensitivity of 77% and a specificity of 55% (LR+ 1.8, 95% CI 1.2–2.5) (figures from a 2023 review; original full text not seen). No independent validation study was found for either. These figures come from the tests as performed and scored in those studies.

Sources: Parvizi J et al. J Hand Surg Br 1998;23:324–7. Bergh TH et al. Emerg Med J 2014;31:659–64; Coventry L et al. Emerg Med J 2023;40:576.