Severe acute respiratory syndrome (SARS) is an infectious disease and some of its symptoms were clinically indistinguishable of those from similar diseases. This study aimed to find the symptom combinations associated with adverse outcome and the therapeutic effects in a cohort of patients with probable SARS retrospectively. In 2003, 123 SARS cases in Beijing were subjected to a strictly western medicine (WM) treatment, or a combined treatment (WM plus Herba houttuyniae injection, addition of individualized herbal treatments when necessary), of which 115 were followed till death or discharge; 8 were transferred and lost to follow-up. In both treatment groups, clinical manifestations were evaluated daily; development of signs and symptoms, and their possible relationship with outcome, were assessed. The relationships between these sign/symptom complexes and outcome under two treatment protocols were evaluated and differences were noted. Dynamic symptom combinations, dividing into the early, the medium-term and the durational symptom clusters, were identified as likely being related to the adverse outcomes of SARS (p < 0.05, p < 0.01). Compared with a strictly WM treatment, the combined treatment resulted in a longer hospital stay (p = 0.028), a non-statistically significant mortality rate decrease (combined treatment: 9.6% versus WM: 11.1%), and a significant improvement of arthralgia and myalgia (p < 0.05) in the early symptom cluster. Additionally, the combined protocol improved arterial oxyhemoglobin saturation significantly at day 22 (p < 0.05). In conclusion, the progress and outcome of SARS may be associated with specific temporal patterns of development in combination of several non-specific signs and symptom complexes, which are also helpful for evaluating the therapeutic effects on SARS patients.