FLEX October 2023

Risk and interval to malignancy of patients with laryngeal dysplasia 367

Potentially relevant papers identified by search n=1037 n =1037

Studies excluded by abstract review n=989 n =989

Papers retrieved for evaluation n=48 n =48

Studies excluded after evaluation (n=39) Not larynx (n=3) Not dysplasia (n=8) Cross sectional studies (n=4) Lack of outcome data (n=6) Minimum follow up not recorded (n=6) Minimum follow up <6 months (n=4) Anomalous MTR, with high risk of bias (n=1) Included patients with early transformation (n=1) Duplication of data (n=4) Review article (n=2) ( n =39) Not larynx ( n =3) Not dysplasia ( n =8) Cross sectional studies ( n =4) Lack of outcome data ( n =6) Minimum follow up not recorded ( n =6) Minimum follow up <6 months ( n =4) ( n =1) transformation ( n =1) Duplication of data ( n =4) Review article ( n =2)

Studies included in meta-analysis n=9 n =9

Fig. 1. Flow diagram showing search strategy and study selection.

Table 2. Description of studies, ordered by minimum follow-up

Mean or median age (years)

Total dysplasia patients

Mean time to transformation (months) 48 (Mild) 38 (Moderate) 25 (Severe) 132 (Severe) 67 (CIS) 29 (Mild) 22 (Moderate) 36 (Severe)

Minimum follow-up (months)

Mean follow-up (months)

Methodology and setting

Percentage follow-up

Study

Country

Ricci 52

89

111

Retrospective hospital (records)

57

Italy

NR

84

Plch 51

Czech

146

60

74

99

Retrospective hospital (records) Retrospective hospital (path) Prospective hospital (records) Retrospective hospital (records)

52

Republic

Sllamniku 53 USA NR

100

317

59

60

Minni 2

Italy

NR

39

NR

100

58

54

Gallo 49

Italy

101

15

173 (Mild) 56 (Moderate) 43 (Severe)

100

116

56

Jeannon 50

UK

72 (median) 15

11 (Combined) 100

113

Retrospective hospital (records)

56

Uno 54

Japan 59

13

60 (Mild) 36 (Severe)

100

26

Retrospective hospital (records ⁄ path)

NR

Blackwell 48

USA 73

12

NR

100

59

Retrospective hospital (records) Retrospective hospital (records)

59

Stenersen 9

Norway 109

12

59 (Severe)

100

41

NR

All studies

940

NR, not recorded; CIS, carcinoma-in-situ.

mation rate (on the transformed scale) against study size (Fig. 2). The expected tendency for larger studies to give central values is seen and there does not appear to be asymmetry or clear evidence of publication bias.

Heterogeneity of studies

The effect of possible bias and heterogeneity between the studies was examined by a funnel plot of transfor

2010 Blackwell Publishing Ltd • Clinical Otolaryngology 35, 364–372