- Email: [email protected]
Preoperative hearing predicts postoperative hearing
Otolaryngology– Head and Neck Surgery DECEMBER 1998
VOLUME 119
NUMBER 6
ORIGINAL ARTICLES Preoperative hearing predicts postoperative hearing BRIAN W. BLAKLEY, MD, PhD, STEVE KIM, MD, and MARY VANCAMP, MS, MPH, Detroit, Michigan
Optimal hearing is one goal of otologic surgery. It is generally presumed that hearing thresholds after tympanoplasty-mastoidectomy surgery are determined by remaining anatomy. We assessed the effect of the disease process on hearing after surgery, and our data do not support this presumption. We studied the relationship between preoperative and postoperative hearing across tympanoplasties with differing anatomies in 124 patients without ossicular reconstruction with prostheses or bone. Our results indicate that poor hearing before surgery is associated with poor hearing after surgery, regardless of anatomy (type of tympanoplasty). This relationship holds for both the airbone gap and the speech-reception threshold. When considering ossicular reconstruction, the surgeon should remember that the disease process itself affects hearing in addition to the physical attributes of the ear. These results should be considered in preoperative counseling, as well as in the interpretation of the otologic literature. (Otolaryngol Head Neck Surg 1998;119:559-63.)
“You can’t make a silk purse from a sow’s ear.” Was this saying intended for ear surgeons? How much hearFrom the Department of Otolaryngology and the Center for Health Care Effectiveness Research, Wayne State University. Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, Washington, D.C., Sept. 29–Oct. 2, 1996. Reprints not available from the authors. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/1/87033
ing improvement can be expected from surgery? In this article we address the question, “Do the ears that start out with poor hearing tend to finish with poorer hearing?” We hypothesize that, for ears with recurrent pus and drainage, hearing results after surgery depend largely on the degree of hearing loss that existed before surgery. Although logical, this notion has not been prominent in the literature. Inasmuch as hearing loss may be a crude measure of the extent of disease, it could be that the worst ears cannot be improved to the level of better-hearing ears. Zollner1 and Wullstein2 pioneered the nomenclature of tympanoplasty. Wullstein’s definitions, later modified by Paparella and Shumrick,3 outlined four types of nonprosthetic tympanoplasty. The basis of this classification was the ossicle or part thereof that was the first to conduct sound in its normal anatomic position. A simplified version of these definitions, as applied in this article, are outlined in Table 1. The tympanoplasty literature contains many articles with a variety of reconstruction techniques. In this article we examine the underlying situation without the variability introduced by reconstruction. The authors of reviews of the results of middle-ear surgery have used anatomic classifications in discussing results4-6 but not preoperative hearing levels. Our reading of the literature regarding tympanoplasty leads us to conclude that the status of the remaining ossicles is considered the main—perhaps even the sole—determinant of postoperative hearing. We were unable to find a single article that systematically considered preoperative hearing as a predictor of postoperative hearing across different types of tympanoplasties. Some authors calculate air-bone gap from a combination of preoperative and postoperative data. This technique does not take 559
560
Otolaryngology– Head and Neck Surgery December 1998
BLAKLEY et al.
Table 1. Anatomic defintion of tympanoplasty
Tympanoplasty type
Definition: Which ossicles conduct sound in their normal anatomic position?
disease causes worse preoperative hearing in patients who require the most anatomically aggressive tympanoplasties. If this is true, then corollary 2 applies. METHODS AND MATERIAL
1 2 3 4
Malleus, incus, stapes Incus, stapes Stapes capitulum Stapes footplate
Table 2.Preoperative and postoperative hearing (dB) Parameter
Preoperative air-bone gap Postoperative air-bone gap Preoperative SRT Postoperative SRT
Mean ± SD
n
19.4 ± 17.4 23.7 ± 17.3 39.3 ± 19.8 45.2 ± 23.0
124 64 124 65
into consideration the relative level of preoperative hearing. Standards for reporting the results of ear surgery with various calculation methods have been proposed,7-9 but none has become the standard. None of these standards takes into account preoperative hearing. Goldenberg and Berliner9 reviewed the results of 550 ossicular reconstructions, applying several criteria to the same data. They found that the definition of “success” (e.g., closure to within 20 or 10 dB) led to significantly different conclusions but that the application of various calculation methods did not. Hearing improvement was not the first goal of the surgical procedures in this study. Classic teaching is that the first goal of tympanomastoidectomy is to control the disease to achieve a dry, safe ear. (See Methods and Material section for definition of “safe.”) If ossicular removal is required, the goal of hearing preservation conflicts with the goal of removing all disease in patients with chronic otitis media, cholesteatoma, or both. If the ossicular anatomy is more important than preoperative hearing in determining postoperative hearing, we expect that these two corollaries would be true: (1) Hearing should be worse with more anatomically aggressive tympanoplasties. (2) The change of hearing levels from before to after surgery does not depend on preoperative hearing level when controlled for type of tympanoplasty. Demonstration of a relationship between type of tympanoplasty and postoperative hearing levels (corollary 1) by itself does not mean that the tympanoplasty is the cause of the hearing loss. It also is possible that the
We sought to assess the auditory effect of tympanoplasty alone, without the possible effects of recurrent disease, ossicular reconstruction or Silastic sheeting in the middle ear. Audiometric data from 166 patients who underwent primary tympanoplasty-mastoidectomy by one of us (BWB) were examined. Audiograms were performed before and approximately 8 weeks after surgery. The protocol was consistent with the Wayne State University policy on human-subjects research. Audiometry We studied the speech-reception threshold (SRT) as a measure of air conduction, as well as its most comparable10 threefrequency–average bone-conduction threshold (average threshold at 500, 1000, and 2000 Hz). We defined the air-bone gap as the three-frequency–average pure-tone air threshold minus the same three-frequency–average bone threshold on the same audiogram. Other authors have used preoperative bone and postoperative air conduction, but this technique makes the air-bone gaps look smaller.11 We defined the change in SRT or air-bone gap as the postoperative value minus the preoperative value. A negative number implies hearing improvement. Surgery Surgery was performed in these ears with the primary goal of removing disease, making the ear dry and safe defined as an epithelialized mastoid cavity without drainage. We defined “disease” as infected granulation tissue from chronic otitis media, cholesterol granuloma, cholesteatoma, or all three. The posterior canal wall was removed if we suspected that all disease had not been removed during simple mastoidectomy and tympanoplasty. For example, if disease was present in the mastoid and epitympanum we presumed that the disease had not been totally eradicated unless the posterior canal wall between them was removed. The tympanoplasty was classified according to the definitions shown in Table 1. We used the following patient-inclusion criteria: (1) diagnosis of cholesteatoma or chronic otitis media (intermittent drainage through a perforation in the tympanic membrane for at least 3 months); (2) only primary surgical procedures were considered; no revision cases were allowed; (3) for chronic otitis media, treatment failure for at least 3 months and two courses of antibiotics; (4) no reconstruction with ossicular prostheses or Silastic in the middle ear was allowed; and (5) audiograms were performed by certified, master’s-level audiologists. When the inclusion criteria were applied, 124 patients
Otolaryngology– Head and Neck Surgery Volume 119 Number 6
BLAKLEY et al.
561
from the original 166 remained in the study. Some postoperative data were not available for some of these patients because of lack of follow-up by the patient, accounting for different sample sizes in the analysis. Statistical Analysis We used ANOVA to test corollary 1. The changes in SRT and air-bone gap across different types of tympanoplasty were assessed. We ran these ANOVAs with and without controlling for preoperative SRT and air-bone gap. Adequate data for 53 patients were available for this analysis. Multiple-regression techniques were used to test corollary 2. The relationships between preoperative and postoperative SRT, as well as preoperative and postoperative air-bone gaps, were assessed. After application of the inclusion criteria, 60 patients had adequate, comparable postoperative data. Missing Data Important information may be lost in missing data. For example, it is possible that patients with better hearing are more likely to comply with follow-up instructions, leading to biased results. We used t tests on preoperative data to assess whether compliant patients with postoperative data had better or worse preoperative hearing than those missing postoperative data. We constrasted the preoperative SRT and air-bone gap for patients with and without postoperative data to determine whether lack of follow-up might be associated with worse hearing. Of the 124 patients who met the inclusion criteria, approximately half were missing postoperative data as a result of poor follow-up. The SPSS/PC+ (Chicago) software package was used for these analyses; significance was set at a p value of 0.05. Levene’s test for equality of variances indicated that the equality-of-variance assumption underlying the ANOVA and t tests was not violated. RESULTS
The preoperative and postoperative air-bone gaps and SRT levels are shown in Table 2. SRT and air-bone gap were slightly worse after surgery, as one might expect without ossicular reconstruction.
Fig. 1. Mean ± SD SRT for different types of tympanoplasty. The difference in mean SRT as more aggressive tympanoplasties (1, least aggressive; 4, most aggressive) were performed is not significant at a p value of 0.05.
Fig. 2. Mean ± SD change in air-bone gap, defined as postoperative value minus preoperative value, for different types of tympanoplasty. No difference in mean airbone gap change was detected as more aggressive tympanoplasties were performed. ANOVA results indicated no significant effect on SRT or air-bone gap according to tympanoplasty type when preoperative SRT and airbone gap were controlled (p = 0.34, n = 60) or when they were not controlled for in this study group.
Statistical Results
Regarding corollary 1, Figures 1 and 2 show the effect of type of tympanoplasty on the change in SRT and change in air-bone gap. ANOVA, when controlled for preoperative SRT and air-bone gap, indicated that the difference in neither SRT nor air-bone gap varied according to the types of tympanoplasty performed (ANOVA; n = 60, p = .34). These conclusions were the same without controlling for preoperative SRT and airbone gap. The results were also the same if the absolute SRT and air-bone gap, instead of the preoperative
minus postoperative difference in these quantities, was considered. These results indicate that—in this set of actively draining ears, at least—the postoperative SRT and airbone gap do not depend on the type of tympanoplasty performed. Corollary 1 is not supported. Regarding corollary 2, multiple-regression results indicated that the postoperative SRT and air-bone gap correlated with preoperative SRT and air-bone gap.
562
Otolaryngology– Head and Neck Surgery December 1998
BLAKLEY et al.
Table 3. Missing-data table: Preoperative hearing Parameter
dB (mean ± SD)
n
p
43.1 ± 18.7 35.5 ± 20.2 23.8 ± 17.3 15.6 ± 17.0
61 63 60 64
0.03
SRT, with postoperative data SRT, missing postoperative data Air-bone gap, with postoperative data Air-bone gap, without postoperative data
Postoperative SRT increased by 0.94 dB for each decibel increase in preoperative SRT (p = .0023). Postoperative air-bone gap increased by 0.48 dB for each increase in preoperative air-bone gap (p < 0.001). A large preoperative SRT or air-bone gap (or both) was associated with a large postoperative SRT or air-bone gap (or both). Postoperative SRT and air-bone gaps covary. Patients with large air-bone gaps are likely to have high SRTs. Our data are consistent with this expectation. These data indicate that postoperative hearing depends largely on preoperative hearing. Corollary 2 is not supported. Considering that our data support neither corollary, we cannot reject the null hypothesis that the remaining ossicular anatomy has no effect on hearing. It seems likely that anatomy plays some role in postsurgical hearing loss, but our data indicate that this is a smaller role than that of preoperative hearing levels. The hypothesis that patients who complied with follow-up instructions had different hearing than those who did not is addressed in Table 3. Using a two-tailed t test, we found significant differences for SRT and airbone gap. As a group, those who did not comply with follow-up instructions had better hearing than those who complied. These findings imply that patients with the worst ears were more likely to seek further care than those with better hearing.
0.008
pertain only to surgery performed for persisting infections with no attempted reconstruction. Ears operated on for the purpose of improving hearing would likely demonstrate different hearing results. Implications for Clinical Practice
In most patients, the goal of attaining a safe ear was achieved within 8 to 12 weeks of surgery. Twelve of the 60 patients required longer follow-up. Five of these 60 patients (8%) either had drainage at the last visit and their treatment was presumed a failure, or they underwent revision surgery. There were no cases of dead ears, facial-nerve or vestibular dysfunction, or other complications in this group. The success rate was 92%.
These data may have several important implications for clinical practice. First, the philosophy of first achieving a safe, dry ear seems sound. A dry, safe ear was attained in 92% of cases. For patients with actively draining ears, our findings imply that hearing loss does not vary significantly for different types of tympanoplasty, so the surgeon should deal aggressively with the disease and not worry about hearing loss. Attempts to “cheat” and preserve anatomy may not result in better hearing and may cause persistence or recurrence of disease. Second, surgeons should anticipate better hearing results in patients with good preoperative hearing. It is conceivable that those with the worst hearing could have the most to gain from surgery. Unfortunately, this does not generally appear true for patients with chronic infection. Patients undergoing reconstructive procedures or stapedectomy may experience different results. Hearing could change with time. Our results reflect the immediate effect of surgery without possible infection, tissue changes, and other time-related effects that could cloud the issue for the future. Although regression analysis showed that hearing did not change over time, a larger population is needed to confidently report these results. Scarring and disease in the middle ear, reflected as hearing loss, may be a crude measure of the severity of disease. Failure to consider factors such as disease severity and preoperative hearing reserve may explain why some results of tympanoplasty are not consistent with the results predicted from acoustical considerations, on the basis of anatomy alone.
DISCUSSION
Implications for the Literature
Our data support the hypothesis that preoperative hearing is an important determinant of postoperative hearing, whereas the anatomy of the tympanoplasty was not. We suspect that this would not have been predicted by most clinicians. We emphasize that these results may
Preoperative hearing should be considered when results of tympanoplasty are being reported in attempts to develop “guidelines” for such reporting. Failure to consider preoperative hearing may explain some of the variability in success rates reported in the literature.
Treatment Results
Otolaryngology– Head and Neck Surgery Volume 119 Number 6
Surgeons operating on patients with better hearing should anticipate better results. Variation of hearing from the predictions of physical, acoustic, or mathematical models (or all three) may be explained in large part by “microanatomic differences” (e.g., scarring, size of resonating chambers, ossicular mobility) rather than macroanatomic differences (e.g., ossicular anatomy). CONCLUSION
In ears with persistent infection, hearing after tympanomastoidectomy surgery depends more on preoperative hearing levels than the nature of the tympanoplasty performed. REFERENCES 1. Zollner F. The principles of plastic surgery of the sound-conducting apparatus. J Laryngol 1955;69:637-52. 2. Wullstein H. Theory and practice of tympanoplasty. Laryngoscope 1956;66:1076-1093.
BLAKLEY et al.
563
3. Paparella MM, Shumrick DA. Otolaryngology. 2nd ed. Vol II. Ear. Philadelphia: Saunders; 1980. p. 1525. 4. Sheehy JL, Shelton C. Tympanoplasty: to stage or not to stage. Otolaryngol Head Neck Surg 1991;104:399-407. 5. Donaldson I. A five-year follow up of incus transposition in relation to the first stage tympanoplasty technique. J Laryngol Otol 1992;106:607-9. 6. Lierle D. Standard classification for surgery of chronic ear infection: the Committee on Conservation of hearing of the American Academy of Ophthalmology and Otolaryngology. Arch Otolaryngol 1956;81:204-5. 7. Monsell EM. Draft guidelines for the evaluation of results of treatment of conductive hearing loss. AAO-HNS Bull 1994;13(4):145. 8. Saki M. Proposal of a guideline in reporting hearing results in middle ear and mastoid surgery. Am J Otol 1994;15:291-3. 9. Goldenberg RA, Berliner KI. Reporting operative hearing results: does choice of outcome measure make a difference? Am J Otol 1995;16:128-34. 10. Jerger JF, Carhart R, Tillman TW, et al. Some relations between normal hearing for pure tones and for speech. J Speech Hear Res 1959;2:126-40. 11. Berliner KI, Doyle KJ, Goldenberg RA. Reporting operative hearing results in stapes surgery: does choice of outcome measure make a difference? Am J Otol 1996;17:521-8.
Congress on the Nose
The 18th International Symposium on Infection and Allergy of the Nose (ISIAN), the 4th Asian Research Symposium on Rhinology (ARSR), and the 3rd ASEAN Rhinology Group Congress (ARC) will be held Feb 14-18, 1999, at the Philippine International Convention Center (PICC), Manila, Philippines. Sponsors are the Philippine Society of Otorhinolaryngology–Head and Neck Surgery (PSO-HNS) and the International Rhinologic Society (IRS). The theme is “Love the Nose.” For further information, contact: Gil M. Vicente, MD, Congress President, Department of ORL-HNS, Jose Reyes Memorial Medical Center, Rizal Avenue Sta. Cruz, Manila, Philippines 1003; phone 632-743-6921; fax, 632-732-1077 or 632-522-0946; e-mail, [email protected]; Website, http://www.psohns.com/isian.
VOLUME 119
NUMBER 6
ORIGINAL ARTICLES Preoperative hearing predicts postoperative hearing BRIAN W. BLAKLEY, MD, PhD, STEVE KIM, MD, and MARY VANCAMP, MS, MPH, Detroit, Michigan
Optimal hearing is one goal of otologic surgery. It is generally presumed that hearing thresholds after tympanoplasty-mastoidectomy surgery are determined by remaining anatomy. We assessed the effect of the disease process on hearing after surgery, and our data do not support this presumption. We studied the relationship between preoperative and postoperative hearing across tympanoplasties with differing anatomies in 124 patients without ossicular reconstruction with prostheses or bone. Our results indicate that poor hearing before surgery is associated with poor hearing after surgery, regardless of anatomy (type of tympanoplasty). This relationship holds for both the airbone gap and the speech-reception threshold. When considering ossicular reconstruction, the surgeon should remember that the disease process itself affects hearing in addition to the physical attributes of the ear. These results should be considered in preoperative counseling, as well as in the interpretation of the otologic literature. (Otolaryngol Head Neck Surg 1998;119:559-63.)
“You can’t make a silk purse from a sow’s ear.” Was this saying intended for ear surgeons? How much hearFrom the Department of Otolaryngology and the Center for Health Care Effectiveness Research, Wayne State University. Presented at the Annual Meeting of the American Academy of Otolaryngology–Head and Neck Surgery, Washington, D.C., Sept. 29–Oct. 2, 1996. Reprints not available from the authors. Copyright © 1998 by the American Academy of Otolaryngology– Head and Neck Surgery Foundation, Inc. 0194-5998/98/$5.00 + 0 23/1/87033
ing improvement can be expected from surgery? In this article we address the question, “Do the ears that start out with poor hearing tend to finish with poorer hearing?” We hypothesize that, for ears with recurrent pus and drainage, hearing results after surgery depend largely on the degree of hearing loss that existed before surgery. Although logical, this notion has not been prominent in the literature. Inasmuch as hearing loss may be a crude measure of the extent of disease, it could be that the worst ears cannot be improved to the level of better-hearing ears. Zollner1 and Wullstein2 pioneered the nomenclature of tympanoplasty. Wullstein’s definitions, later modified by Paparella and Shumrick,3 outlined four types of nonprosthetic tympanoplasty. The basis of this classification was the ossicle or part thereof that was the first to conduct sound in its normal anatomic position. A simplified version of these definitions, as applied in this article, are outlined in Table 1. The tympanoplasty literature contains many articles with a variety of reconstruction techniques. In this article we examine the underlying situation without the variability introduced by reconstruction. The authors of reviews of the results of middle-ear surgery have used anatomic classifications in discussing results4-6 but not preoperative hearing levels. Our reading of the literature regarding tympanoplasty leads us to conclude that the status of the remaining ossicles is considered the main—perhaps even the sole—determinant of postoperative hearing. We were unable to find a single article that systematically considered preoperative hearing as a predictor of postoperative hearing across different types of tympanoplasties. Some authors calculate air-bone gap from a combination of preoperative and postoperative data. This technique does not take 559
560
Otolaryngology– Head and Neck Surgery December 1998
BLAKLEY et al.
Table 1. Anatomic defintion of tympanoplasty
Tympanoplasty type
Definition: Which ossicles conduct sound in their normal anatomic position?
disease causes worse preoperative hearing in patients who require the most anatomically aggressive tympanoplasties. If this is true, then corollary 2 applies. METHODS AND MATERIAL
1 2 3 4
Malleus, incus, stapes Incus, stapes Stapes capitulum Stapes footplate
Table 2.Preoperative and postoperative hearing (dB) Parameter
Preoperative air-bone gap Postoperative air-bone gap Preoperative SRT Postoperative SRT
Mean ± SD
n
19.4 ± 17.4 23.7 ± 17.3 39.3 ± 19.8 45.2 ± 23.0
124 64 124 65
into consideration the relative level of preoperative hearing. Standards for reporting the results of ear surgery with various calculation methods have been proposed,7-9 but none has become the standard. None of these standards takes into account preoperative hearing. Goldenberg and Berliner9 reviewed the results of 550 ossicular reconstructions, applying several criteria to the same data. They found that the definition of “success” (e.g., closure to within 20 or 10 dB) led to significantly different conclusions but that the application of various calculation methods did not. Hearing improvement was not the first goal of the surgical procedures in this study. Classic teaching is that the first goal of tympanomastoidectomy is to control the disease to achieve a dry, safe ear. (See Methods and Material section for definition of “safe.”) If ossicular removal is required, the goal of hearing preservation conflicts with the goal of removing all disease in patients with chronic otitis media, cholesteatoma, or both. If the ossicular anatomy is more important than preoperative hearing in determining postoperative hearing, we expect that these two corollaries would be true: (1) Hearing should be worse with more anatomically aggressive tympanoplasties. (2) The change of hearing levels from before to after surgery does not depend on preoperative hearing level when controlled for type of tympanoplasty. Demonstration of a relationship between type of tympanoplasty and postoperative hearing levels (corollary 1) by itself does not mean that the tympanoplasty is the cause of the hearing loss. It also is possible that the
We sought to assess the auditory effect of tympanoplasty alone, without the possible effects of recurrent disease, ossicular reconstruction or Silastic sheeting in the middle ear. Audiometric data from 166 patients who underwent primary tympanoplasty-mastoidectomy by one of us (BWB) were examined. Audiograms were performed before and approximately 8 weeks after surgery. The protocol was consistent with the Wayne State University policy on human-subjects research. Audiometry We studied the speech-reception threshold (SRT) as a measure of air conduction, as well as its most comparable10 threefrequency–average bone-conduction threshold (average threshold at 500, 1000, and 2000 Hz). We defined the air-bone gap as the three-frequency–average pure-tone air threshold minus the same three-frequency–average bone threshold on the same audiogram. Other authors have used preoperative bone and postoperative air conduction, but this technique makes the air-bone gaps look smaller.11 We defined the change in SRT or air-bone gap as the postoperative value minus the preoperative value. A negative number implies hearing improvement. Surgery Surgery was performed in these ears with the primary goal of removing disease, making the ear dry and safe defined as an epithelialized mastoid cavity without drainage. We defined “disease” as infected granulation tissue from chronic otitis media, cholesterol granuloma, cholesteatoma, or all three. The posterior canal wall was removed if we suspected that all disease had not been removed during simple mastoidectomy and tympanoplasty. For example, if disease was present in the mastoid and epitympanum we presumed that the disease had not been totally eradicated unless the posterior canal wall between them was removed. The tympanoplasty was classified according to the definitions shown in Table 1. We used the following patient-inclusion criteria: (1) diagnosis of cholesteatoma or chronic otitis media (intermittent drainage through a perforation in the tympanic membrane for at least 3 months); (2) only primary surgical procedures were considered; no revision cases were allowed; (3) for chronic otitis media, treatment failure for at least 3 months and two courses of antibiotics; (4) no reconstruction with ossicular prostheses or Silastic in the middle ear was allowed; and (5) audiograms were performed by certified, master’s-level audiologists. When the inclusion criteria were applied, 124 patients
Otolaryngology– Head and Neck Surgery Volume 119 Number 6
BLAKLEY et al.
561
from the original 166 remained in the study. Some postoperative data were not available for some of these patients because of lack of follow-up by the patient, accounting for different sample sizes in the analysis. Statistical Analysis We used ANOVA to test corollary 1. The changes in SRT and air-bone gap across different types of tympanoplasty were assessed. We ran these ANOVAs with and without controlling for preoperative SRT and air-bone gap. Adequate data for 53 patients were available for this analysis. Multiple-regression techniques were used to test corollary 2. The relationships between preoperative and postoperative SRT, as well as preoperative and postoperative air-bone gaps, were assessed. After application of the inclusion criteria, 60 patients had adequate, comparable postoperative data. Missing Data Important information may be lost in missing data. For example, it is possible that patients with better hearing are more likely to comply with follow-up instructions, leading to biased results. We used t tests on preoperative data to assess whether compliant patients with postoperative data had better or worse preoperative hearing than those missing postoperative data. We constrasted the preoperative SRT and air-bone gap for patients with and without postoperative data to determine whether lack of follow-up might be associated with worse hearing. Of the 124 patients who met the inclusion criteria, approximately half were missing postoperative data as a result of poor follow-up. The SPSS/PC+ (Chicago) software package was used for these analyses; significance was set at a p value of 0.05. Levene’s test for equality of variances indicated that the equality-of-variance assumption underlying the ANOVA and t tests was not violated. RESULTS
The preoperative and postoperative air-bone gaps and SRT levels are shown in Table 2. SRT and air-bone gap were slightly worse after surgery, as one might expect without ossicular reconstruction.
Fig. 1. Mean ± SD SRT for different types of tympanoplasty. The difference in mean SRT as more aggressive tympanoplasties (1, least aggressive; 4, most aggressive) were performed is not significant at a p value of 0.05.
Fig. 2. Mean ± SD change in air-bone gap, defined as postoperative value minus preoperative value, for different types of tympanoplasty. No difference in mean airbone gap change was detected as more aggressive tympanoplasties were performed. ANOVA results indicated no significant effect on SRT or air-bone gap according to tympanoplasty type when preoperative SRT and airbone gap were controlled (p = 0.34, n = 60) or when they were not controlled for in this study group.
Statistical Results
Regarding corollary 1, Figures 1 and 2 show the effect of type of tympanoplasty on the change in SRT and change in air-bone gap. ANOVA, when controlled for preoperative SRT and air-bone gap, indicated that the difference in neither SRT nor air-bone gap varied according to the types of tympanoplasty performed (ANOVA; n = 60, p = .34). These conclusions were the same without controlling for preoperative SRT and airbone gap. The results were also the same if the absolute SRT and air-bone gap, instead of the preoperative
minus postoperative difference in these quantities, was considered. These results indicate that—in this set of actively draining ears, at least—the postoperative SRT and airbone gap do not depend on the type of tympanoplasty performed. Corollary 1 is not supported. Regarding corollary 2, multiple-regression results indicated that the postoperative SRT and air-bone gap correlated with preoperative SRT and air-bone gap.
562
Otolaryngology– Head and Neck Surgery December 1998
BLAKLEY et al.
Table 3. Missing-data table: Preoperative hearing Parameter
dB (mean ± SD)
n
p
43.1 ± 18.7 35.5 ± 20.2 23.8 ± 17.3 15.6 ± 17.0
61 63 60 64
0.03
SRT, with postoperative data SRT, missing postoperative data Air-bone gap, with postoperative data Air-bone gap, without postoperative data
Postoperative SRT increased by 0.94 dB for each decibel increase in preoperative SRT (p = .0023). Postoperative air-bone gap increased by 0.48 dB for each increase in preoperative air-bone gap (p < 0.001). A large preoperative SRT or air-bone gap (or both) was associated with a large postoperative SRT or air-bone gap (or both). Postoperative SRT and air-bone gaps covary. Patients with large air-bone gaps are likely to have high SRTs. Our data are consistent with this expectation. These data indicate that postoperative hearing depends largely on preoperative hearing. Corollary 2 is not supported. Considering that our data support neither corollary, we cannot reject the null hypothesis that the remaining ossicular anatomy has no effect on hearing. It seems likely that anatomy plays some role in postsurgical hearing loss, but our data indicate that this is a smaller role than that of preoperative hearing levels. The hypothesis that patients who complied with follow-up instructions had different hearing than those who did not is addressed in Table 3. Using a two-tailed t test, we found significant differences for SRT and airbone gap. As a group, those who did not comply with follow-up instructions had better hearing than those who complied. These findings imply that patients with the worst ears were more likely to seek further care than those with better hearing.
0.008
pertain only to surgery performed for persisting infections with no attempted reconstruction. Ears operated on for the purpose of improving hearing would likely demonstrate different hearing results. Implications for Clinical Practice
In most patients, the goal of attaining a safe ear was achieved within 8 to 12 weeks of surgery. Twelve of the 60 patients required longer follow-up. Five of these 60 patients (8%) either had drainage at the last visit and their treatment was presumed a failure, or they underwent revision surgery. There were no cases of dead ears, facial-nerve or vestibular dysfunction, or other complications in this group. The success rate was 92%.
These data may have several important implications for clinical practice. First, the philosophy of first achieving a safe, dry ear seems sound. A dry, safe ear was attained in 92% of cases. For patients with actively draining ears, our findings imply that hearing loss does not vary significantly for different types of tympanoplasty, so the surgeon should deal aggressively with the disease and not worry about hearing loss. Attempts to “cheat” and preserve anatomy may not result in better hearing and may cause persistence or recurrence of disease. Second, surgeons should anticipate better hearing results in patients with good preoperative hearing. It is conceivable that those with the worst hearing could have the most to gain from surgery. Unfortunately, this does not generally appear true for patients with chronic infection. Patients undergoing reconstructive procedures or stapedectomy may experience different results. Hearing could change with time. Our results reflect the immediate effect of surgery without possible infection, tissue changes, and other time-related effects that could cloud the issue for the future. Although regression analysis showed that hearing did not change over time, a larger population is needed to confidently report these results. Scarring and disease in the middle ear, reflected as hearing loss, may be a crude measure of the severity of disease. Failure to consider factors such as disease severity and preoperative hearing reserve may explain why some results of tympanoplasty are not consistent with the results predicted from acoustical considerations, on the basis of anatomy alone.
DISCUSSION
Implications for the Literature
Our data support the hypothesis that preoperative hearing is an important determinant of postoperative hearing, whereas the anatomy of the tympanoplasty was not. We suspect that this would not have been predicted by most clinicians. We emphasize that these results may
Preoperative hearing should be considered when results of tympanoplasty are being reported in attempts to develop “guidelines” for such reporting. Failure to consider preoperative hearing may explain some of the variability in success rates reported in the literature.
Treatment Results
Otolaryngology– Head and Neck Surgery Volume 119 Number 6
Surgeons operating on patients with better hearing should anticipate better results. Variation of hearing from the predictions of physical, acoustic, or mathematical models (or all three) may be explained in large part by “microanatomic differences” (e.g., scarring, size of resonating chambers, ossicular mobility) rather than macroanatomic differences (e.g., ossicular anatomy). CONCLUSION
In ears with persistent infection, hearing after tympanomastoidectomy surgery depends more on preoperative hearing levels than the nature of the tympanoplasty performed. REFERENCES 1. Zollner F. The principles of plastic surgery of the sound-conducting apparatus. J Laryngol 1955;69:637-52. 2. Wullstein H. Theory and practice of tympanoplasty. Laryngoscope 1956;66:1076-1093.
BLAKLEY et al.
563
3. Paparella MM, Shumrick DA. Otolaryngology. 2nd ed. Vol II. Ear. Philadelphia: Saunders; 1980. p. 1525. 4. Sheehy JL, Shelton C. Tympanoplasty: to stage or not to stage. Otolaryngol Head Neck Surg 1991;104:399-407. 5. Donaldson I. A five-year follow up of incus transposition in relation to the first stage tympanoplasty technique. J Laryngol Otol 1992;106:607-9. 6. Lierle D. Standard classification for surgery of chronic ear infection: the Committee on Conservation of hearing of the American Academy of Ophthalmology and Otolaryngology. Arch Otolaryngol 1956;81:204-5. 7. Monsell EM. Draft guidelines for the evaluation of results of treatment of conductive hearing loss. AAO-HNS Bull 1994;13(4):145. 8. Saki M. Proposal of a guideline in reporting hearing results in middle ear and mastoid surgery. Am J Otol 1994;15:291-3. 9. Goldenberg RA, Berliner KI. Reporting operative hearing results: does choice of outcome measure make a difference? Am J Otol 1995;16:128-34. 10. Jerger JF, Carhart R, Tillman TW, et al. Some relations between normal hearing for pure tones and for speech. J Speech Hear Res 1959;2:126-40. 11. Berliner KI, Doyle KJ, Goldenberg RA. Reporting operative hearing results in stapes surgery: does choice of outcome measure make a difference? Am J Otol 1996;17:521-8.
Congress on the Nose
The 18th International Symposium on Infection and Allergy of the Nose (ISIAN), the 4th Asian Research Symposium on Rhinology (ARSR), and the 3rd ASEAN Rhinology Group Congress (ARC) will be held Feb 14-18, 1999, at the Philippine International Convention Center (PICC), Manila, Philippines. Sponsors are the Philippine Society of Otorhinolaryngology–Head and Neck Surgery (PSO-HNS) and the International Rhinologic Society (IRS). The theme is “Love the Nose.” For further information, contact: Gil M. Vicente, MD, Congress President, Department of ORL-HNS, Jose Reyes Memorial Medical Center, Rizal Avenue Sta. Cruz, Manila, Philippines 1003; phone 632-743-6921; fax, 632-732-1077 or 632-522-0946; e-mail, [email protected]; Website, http://www.psohns.com/isian.