3.3. Study characteristics and network geometry
In total, 960 subjects were randomly assigned to pharmacological intervention and 751 to placebo. The included RCTs were published from 2002 to 2020 and involved orthopedic(n=2), urinary(n=4), abdominal (n=10), gynecologic (n=9), thyroid surgery (n=5), thoracic(n=1) and ear-nose-throat surgery (n=2). The basic characteristics of enrolled studies are described in Table 1 .
The network geometry plot (Figure 2 ) shows the network of eligible comparisons for postoperative pain intensity at rest at 24h (A) and the incidence of PONV (B). Of the 33 enrolled studies, 28 studies involving 20 treatments reported the postoperative pain intensity at rest at 24h after surgery, and 27 studies involving 17 treatments reported the incidence of PONV. All treatments had at least one placebo-controlled trial. Each treatment was represented by a node and linked by an edge when a direct comparison was performed. More sample sizes, bigger node. Similarly, more studies involved, thicker edge.