Background: Cervical strength and stability is often addressed in rehabilitation of people with neck pain. These functions may be associated with emergence and retention of neck pain in cases where the neck strength does not meet daily demands. However, clear empiric support for these notions are lacking, partly due to a scarcity of well controlled dynamometry studies with large samples. First, clarification is needed whether neck strength is in fact reduced in people with neck pain and to resolve the diagnostic performance of neck strength tests.

Purpose: To compare neck muscle strength of women with non-specific long-term neck pain and healthy controls. Also, to assess the diagnostic performance of neck strength tests by assessing their discriminative ability to discern women with and without neck pain.

Methods: The study had a cross-sectional design with data on cervical strength derived from the baseline measurement of a randomized controlled trial (RCT). 80 women with long-term non-specific neck pain (NP) were compared with 40 healthy women (CON). The NP group was a subsample from the RCT selected so that there were no group differences (NP-CON) for body weight and physical activity. Cervical strength assessment included dynamometry of cervico-thoracic extension (CTE) and flexion (CTF) in sitting and cranio-cervical flexion (CCF) in standing, all performed with isometric maximum voluntary contraction (iMVC). Independent samples Mann-Whitney U test and T-test were used to assess group differences for iMVC of the tests. Diagnostic accuracy was further assessed with a receiver operating characteristic (ROC) curve by plotting the true positive rate (sensitivity) as a function of the false positive rate (1 - specificity). The area under the ROC curve (AUC) with 95% confidence interval was used to determine discriminative ability of the tests. The optimal cut-off value to discern NP from CON with corresponding sensitivity and specificity was also determined.

Results: Women with neck pain produced significant lower iMVC in CTE (28%), CTF (26%) and CCF (33%) (all p< 0.001). The ability of CTE, CTF and CCF to discriminate between NP and CON showed moderate accuracy (AUC 0.83, 0.78 and 0.73, respectively). The cut-off value of 165.7 N in CTE had a sensitivity of 0.725 and a specificity of 0.8. The corresponding values for CTF and CCF were 85.8 N (sensitivity 0.8; specificity 0.692) and 4.2 Nm (sensitivity 0.575; specificity 0.9).

Conclusion(s): The neck pain group had less neck muscle strength than controls in all tests. The diagnostic performance of the neck strength tests, judged as discriminative ability to discern neck pain from control participants, were moderate. These results support earlier findings of reduced neck strength in people with neck pain. Also, neck strength tests may have a complementary value in the assessment of neck pain persons.

Implications: The results highlights that impaired neck strength, both in global cervico-thoracic and deep cranio-cervical muscles, is most likely a characteristic of people with long-term neck pain. The study also shows that strength tests could be used with fairly good discriminative precision and may thereby be valuable assessment tools.