I recently submitted an article to a journal which included the floowing information, and got criticism from one of the reviwers for not having controlled for covariates in the analysis. Is there really a need for multivariate analysis when all the other independent variables, but one have been found to be unrelated with the dependent variable.

Variable
All tAb (+) patients tAb (-) patients p


Age 37.4±11.4 39.9±10.5 p*=0.478

Gender 6/5 40/23 p=0.730

Age of disease onset

23.5±8.2 24.4±7.6 p=0.710

Total BPRS rating

19.1±5.3 19.7±9.4 p=0.820


Current use of an antidepressant agent f (%)

1 (9) 10 (15) p=0.560

Current use of a mood stabilizer g (%)

1 (9) 8 (12) p=0.736

Serum prolactin concentration (ng/dL)

39.8 19.9 p=0.039