Table 6 Multivariable Cox regression analysis for associations to definitive voice prosthesis removal*

From: Oncological and functional outcome after laryngectomy for laryngeal and hypopharyngeal cancer: a population-based analysis in Germany from 2001 to 2020

Parameter

HR

95%-CI

p

Model 1: inclusion of all significant factors from the univariate analysis

Postoperative adjuvant treatment

 No

1

Reference

 

 Radiotherapy

1.140

0.372–3.489

0.819

 Radiochemotherapy

0.903

0.239–3.406

0.880

 Chemotherapy

2.631

0.275–25.165

0.401

Number of prosthesis changes

 > Median 2 changes

1

Reference

 

 < Median 2 changes

2.481

1.063–5.786

0.035

Postoperative aspiration

 No

1

Reference

 

 Yes

2.524

0.839–7.592

0.099

Disturbed wound healing

 No

1

Reference

 

 Yes

1.461

0.561–3.800

0.437

Head neck swelling, needing treatment

 No

1

Reference

 

 Yes

1.871

0.714–4.904

0.202

Model 2: like model 1 but omitting the number of voice prosthesis changes as permanent removal might be directly related to a lower number of changes

Postoperative adjuvant treatment

 No

1

Reference

 

 Radiotherapy

1.031

0.374–2.837

0.954

 Radiochemotherapy

0.813

0.247–2.677

0.734

 Chemotherapy

3.947

0.892–17,464

0.070

Postoperative aspiration

 No

1

Reference

 

 Yes

2.679

1.001–7.167

0.050

Disturbed wound healing

 No

1

  

 Yes

1.398

0.598–3.268

0.439

Head neck swelling, needing treatment

 No

1

Reference

 

 Yes

2.218

0.967–5.087

0.060

Model 3: like model 2 but omitting the factor postoperative adjuvant treatment as only 2.4% of the patients received a postoperative chemotherapy as monotherapy

Postoperative aspiration

 No

1

Reference

 

 Yes

3.126

1.207–8.095

0.019

Disturbed wound healing

 No

1

Reference

 

 Yes

1.480

0.632–3.468

0.366

Head neck swelling, needing treatment

 No

1

Reference

 

 Yes

2.268

0.985–5.227

0.054

  1. Only significant parameters from the univariate analyses were included in the multivariate analyses.
  2. Significant values are in bold.
Back to article page