Table 3. Clinical value of symptoms and signs in patients presenting with rectal bleeding in terms of colorectal cancer.
| No of studiesa | No of patients | Sens | (95% CI) | Spec | (95% CI) | Pooled PLR | (95% CI) | |
|---|---|---|---|---|---|---|---|---|
| Patient characteristics | ||||||||
| Male | 5 | 1253 | 0.58 | (0.48–0.67) | 0.52 | (0.48–0.56) | 1.21 | (1.00–1.46) |
| Age <40 yearsb | 2 | 745 | 0.03 | (0.00–0.16) | 0.73 | (0.69–0.76) | 0.32 | (0.05–2.21) |
| Age 40–59 yearsb | 4 | 1387 | 0.09 | (0.04–0.19) | 0.79 | (0.70–0.86) | 0.41 | (0.18–0.90) |
| Age ⩾ 60 yearsb | 6 | 1760 | 0.66 | (0.45–0.83) | 0.76 | (0.68–0.83) | 2.79 | (2.00–3.90) |
| Family history colorectal cancer | 3 | 886 | 0.15 | (0.06–0.28) | 0.85 | (0.82–0.87) | 1.05 | (0.16–6.88) |
| Symptoms | ||||||||
| Dark red bloodc | 4 | 949 | 0.22 | (0.13–0.34) | 0.84 | (0.69–0.93) | 1.37 | (0.59–3.30) |
| Weight loss | 7 | 1737 | 0.17 | (0.06–0.37) | 0.91 | (0.83–0.96) | 1.89 | (1.03–3.07) |
| Abdominal pain | 7 | 1739 | 0.25 | (0.04–0.62) | 0.73 | (0.52–0.89) | 0.94 | (0.19–1.59) |
| Changed bowel habit | 5 | 1254 | 0.62 | (0.18–0.94) | 0.68 | (0.53–0.80) | 1.92 | (0.54–3.57) |
| Blood mixed with the stool | 5 | 1225 | 0.40 | (0.04–0.93) | 0.81 | (0.23–0.98) | 1.91 | (0.75–5.51) |
| Previous history of rectal bleedingd | 2 | 425 | 0.30 | (0.05–0.41) | 0.66 | (0.63–0.71) | 0.58 | (0.14–1.41) |
| Perianal symptoms – pain on defecation | 2 | 411 | 0.22 | (0.13–0.36) | 0.41 | (0.22–0.78) | 0.49 | (0.25–0.97) |
| Perianal symptoms – itch/eczema | 2 | 414 | 0.17 | (0.07–0.33) | 0.87 | (0.73–0.95) | 1.31 | (0.25–6.21) |
| Signs and diagnostic tests | ||||||||
| Rectal palpation – haemorrhoid | 2 | 354 | 0.24 | (0.09–0.45) | 0.73 | (0.46–0.91) | 0.51 | (0.09–2.97) |
| Anaemia (Hb ♀<12.0 g per 100 ml ♂<13.3 g per 100 ml) | 2 | 700 | 0.17 | (0.05–0.35) | 0.95 | (0.93–0.96) | 3.67 | (1.30–10.35) |
Abbreviations: CI=confidence interval; Hb, haemoglobin; PLR=positive likelihood ratio.
a
Norrelund and Norrelund (1996) consists of two independent sub-studies, and therefore are independently assessed. In the column ‘no of studies’ these two substudies are counted as two separate studies.
b
There is a slight age overlap between the individual studies.
c
The reference category of dark red blood consists of patients having bright red blood or a colour in between.
d
The reference category of previous history of rectal bleeding consists of patients having a first episode of rectal bleeding.