Skip to main content

Table 1 Summary of the two included studies

From: Cilostazol for the prevention of pneumonia: a systematic review

Study and country Yamaya et al. 2001 [16], Japan Shinohara et al. 2006 [10], Japan
Population History of cerebral infarction but not bedridden History of cerebral infarction, age < 80 years
Mean age Mean age: 76.5 ± 2.1 years Mean age: 65.0 ± 8.7 years
Gender No statement about gender Female 34.4%
Risk factors No statement about risk factors for pneumonia No statement about risk factors for pneumonia
Criteria for diagnosis of pneumonia New pulmonary infiltrate seen on chest radiographs (assessed by two radiologists not involved in the study) and one of the following: cough, body temperature > 37.8 °C, subjective dyspnea Description by attending physician in the medical chart, chest radiographs and blood test
Intervention and control (n, allocated) I: cilostazol 100 mg per day (n: not stated) I: cilostazol 200 mg per day (n: 547)
C: no active treatment (n: not stated) C: placebo (n: 548)
328 patients were allocated in total but the number of patients in each group at allocation was not reported  
Intervention period 3 years 3.3 years on average
Number of patients analyzed I = 125, C = 145, see comments I = 531, C = 533, see comments
Outcomes Pneumonia Pneumonia
I: 12 (9.6%), C: 35 (24.1%) I: 3 (0.56%), C: 15 (2.8%)
RR = 0.40 (95% CI 0.22–0.73)** RR = 0.20 (95% CI 0.06–0.69)**
Cerebral infarction Cerebral infarction [17]
I: 9 (7.2%), C: 24 (16.6%) I: 30 (5.6%), C: 57 (10.7%)
RR = 0.43 (95% CI 0.21–0.90)* RR = 0.53 (95% CI 0.34–0.81)*
Bleeding Bleeding [17]
I: 12 (9.6%), severity of the bleeding events was not reported Fatal intracranial hemorrhage
C: No statement about adverse events I: 0, C: 1
  Nonfatal intracranial hemorrhage
  I: 4 (0.75%), C: 6 (1.1%)
  Other bleeding
  I: 15a (2.8%), C: 11 (2.1%)
Risk of bias
Random sequence generation Low Low
Allocation concealment Unclear: Inadequate description Low
Blinding of participants and personnel High: Open-label trial Low
Blinding of outcome assessment Low Unclear: Post-hoc analysis of a RCT [17]
Incomplete outcome data High: As-treated analysis Unclear: Reasons for dropouts were unclear
Selective reporting High: No protocol available, No statement about adverse events in the control group High: Post-hoc analysis of a RCT [17]
Other bias Unclear: No statement about funding Unclear: No statement about funding but original RCT [17] was supported by Otsuka Pharmaceutical
Comments Reported in a Letter to the Editor 1095 patients were randomized but 31 patients were excluded from analysis because of “serious protocol violations”
328 patients were randomized but 58 patients were excluded from analysis; 31 patients died from causes other than pneumonia; additionally, 27 patients in cilostazol group dropped out because of adverse events Of 1064 patients, about half of the patients discontinued the treatment (I: 54%, C: 49%)
  1. C control, CI confidence interval, I intervention, RR risk ratio
  2. *P < 0.05, **P < 0.01
  3. a Including one gastrointestinal bleeding