Review Article

Observational Studies on Evaluating the Safety and Adverse Effects of Traditional Chinese Medicine

Table 2

Number of new cases ( ), adjusted hazards ratios (HR), and 95% confidence intervals (CI) estimated from multivariate Cox regression model for chronic kidney disease and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) for new occurrence of kidney failure and urinary tract cancer from multivariable logistic regression models on a random sample of the National Health Insurance database followed from 1997–2002.
(a)

Cox regression model for chronic kidney diseasea
Study Chinese herbNo. of casesHR95% CI

Mu-Tong
 0 g  1,9791.0
 1–30 g 248 1.00.8–1.1
 31–60 g 631.31.03–1.8
 61–100 g 271.40.96–2.1
 101–200 g 221.71.1–2.6
 >200 g 40.7 0.3–1.9
Fangchi
 0 g 1,875 1.0
 1–30 g 389 1.0 0.9–1.2
 31–60 g 42 1.3 0.98–1.9
 61–100 g 18 1.8 1.1–2.8
 101–200 g 10 1.4 0.8–2.7
 >200 g 9 2.2 1.1–4.2

(b)

Multiple logistic regression model for kidney failureb
Study Chinese herbNo. of cases/controlsAdjusted OR95% CI

Mu-Tong
 0 g22,188/157,9391.0
 1–30 g2,542/20,122 1.12 0.86–1.47
 31–60 g492/3,729 1.16 0.83–1.62
 61–100 g226/1,569 1.47 1.01–2.14
 101–200 g209/1,054 2.14 1.47–3.11
 >200 g 186/438 5.82 3.89–8.71
Fangchi
 0 g21,985/157,5431.0
 1–30 g3,145/24,8680.68 0.58–0.78
 31–60 g362/1,5281.14 0.91–1.44
 61–100 g169/4921.60 1.20–2.14
 101–200 g116/2951.62 1.17–2.23
 >200 g 66/1251.94 1.29–2.92

(c)

Multiple logistic regression model for urinary tract cancerc
Study Chinese herbNo. of cases/controlsAdjusted OR 95% CI

Mu-Tong
 0 g3,987/149,4641.0
 1–60 g489/22,354 1.0 0.9–1.2
 61–100 g50/1,485 1.6 1.3–2.1
 101–200 g 46/1003 2.0 1.4–2.7
 >200 g 22/395 2.1 1.3–3.4
Fangchi
 0 g 3,927/150,456 1.0
 1–60 g 623/23,456 0.9 0.8–1.0
 61–100 g15/427 0.7 0.4–1.2
 >100 g29/362 1.3 0.9–2.0
Estimated cumulative dose of aristolochic acid, in mg
 03,274/121,8201.0
 1–1501151/48,8691.0 0.96–1.1
 151–25069/2,0321.4 1.1–1.8
 251–50064/1,4031.6 1.2–2.1
 >50036/5772.0 1.4–2.9

Hazards ratios (HR) and 95% confidence intervals (CI) obtained from Cox proportional hazards regression models with all variables (sex, age, hypertension, diabetes, cumulative dosage of nonsteroidal antiinflammatory drugs, Mu Xiang, Mu-Tong, and Fangchi) fitted simultaneously.
bMultivariable odds ratios were adjusted for potential confounders (sex, age, hypertension, diabetes, chronic hepatitis, chronic urinary tract infection, chronic neuralgia, musculoskeletal disease, cumulative dosage of nonsteroidal antiinflammatory drugs, Mu Xiang, Mu-Tong, and Fangchi) fitted simultaneously.
cMultivariable odds ratios were adjusted for potential confounders (sex, age, hypertension, diabetes, residence in township where black foot disease was endemic, chronic urinary tract infection, Mu Xiang, Mu-Tong, and Fangchi) fitted simultaneously.