Small or inconsistent changes in values do not necessarily indicate a clinically important effect and should be confirmed by repeat testing before being otherwise investigated.