The Effect of Hamulus Fracture on the Outcome of Palatoplasty: A Preliminary Report of a Prospective, Alternating Study

ALEX A. KANE, LUN-JOU LO, BEN-DONG YEN, YU-RAY CHEN, M. SAMUEL NOORDHOFF,
Objective: To determine whether, in performing palatoplasty, fracture of the pterygoid hamulus is benecial, detrimental, or neutral with respect to intraoperative and perioperative complications, hearing outcome, and speech outcome. Design: Prospective, alternating. Setting: Institutional, tertiary cleft palate center, Chang Gung Memorial Hospital, Taipei, Taiwan. Participants: A total of 173 patients enrolled in the study, of whom 161 had charts available for analysis. Interventions: During the performance of palatoplasty, 85 patients received hamulus fracture and 76 patients did not. All palatoplasties were performed by the same surgeon. Main Outcome Measures: (1) Surgical outcomes, including patient demographic data, palatoplasty type and duration, blood loss, incidences of oronasal stulae, temporary mucosal dehiscence, and postoperative bleeding; (2) otolaryngological outcomes, including hearing results as judged by auditory brainstem response testing, myringotomy tube data describing rates of tube extrusion, and culture results from sampled effusions; and (3) preliminary speech outcomes as described by judgments of overall velopharyngeal function from perceptual speech samples. Results: No statistically signicant differences in any of the measured surgical, otolaryngological, or preliminary speech outcomes were found between the groups who did and did not receive hamulus fracture. Conclusions: On the basis of these results, we are unable to advocate the performance of hamulus fracture as an operative maneuver during the performance of primary palatoplasty. The historical rationale and theoretical advantage of this maneuver have not been demonstrated here nor have any detrimental effects of the maneuver been measured.
KEY WORDS: cleft palate, palatoplasty, pterygoid hamulus

M.D. M.D. M.D. M.D. M.D.

Palatoplasty, the surgical repair of cleft palate, consists of a series of technical maneuvers that an individual surgeon chooses from a repertoire developed over the past 150 years. The cleft surgeon includes some maneuvers and excludes others, for a variety of reasons, which might include hard and reliable outcome data, impressions about outcome, opinions of the surgeons educators, opinions of other professionals, and the surDr. Alex A. Kane is Fellow, Dr. Lun-Jou Lo is Associate Professor, and Dr. Yu-Ray Chen and Dr. M. Samuel Noordhoff are Professors, Department of Plastic and Reconstructive Surgery. Dr. Ben-Dong Yen is Assistant Professor, Department of Otorhinolaryngology, Chang Gung Memorial Hospital, Taipei, Taiwan. Submitted May 1999; Accepted December 1999. Reprint requests: Lun-Jou Lo, M.D., Department of Plastic and Reconstructive Surgery, Chang Gung Memorial Hospital, 199 Tun Hwa North Road, Taipei, Taiwan 105. E-mail lunjoulo@ms1.hinet.net. 506

geons personal preferences. One of the rst ancillary procedures for palatoplasty, beyond mucosal repair, was hamulus fracture. This maneuver was advocated by Billroth (1889) to minimize postoperative palatal dehiscence. Whether hamulus fracture has any effect upon palatal integrity or any other aspect of cleft palate care has not been rigorously assessed. Among the most important goals of reconstructing the cleft palate are the restoration of normal anatomic separation of the oral and nasal cavities, normal speech, and maintenance of ear function. The incidences of middle ear effusion, otitis media, and hearing impairment are high in patients with cleft palate (Stool and Randall, 1967; Yules, 1970; Paradise, 1975; Chaudhuri and Bowen-Jones, 1978; Gopalakrishna et al., 1984; Fria et al., 1987). This is thought to be due to malfunction of the tensor veli palatini. The tensor is attached to the skull base and the eustachian tube superiorly, descending inferiorly and

Kane et al., EFFECTS OF HAMULUS FRACTURE ON PALATOPLASTY OUTCOME

507

medially, forming a tendonous structure that hooks around the hamulus, joining the palatal aponeurosis (Ross, 1971; Bluestone et al., 1975; Shibahara and Sando, 1988; Spauwen et al., 1991; Barsoumian et al., 1998). During palate repair, some surgeons fracture the hamulus processes and dislocate the tensor muscle away from the process in order to minimize tension on the repair and thereby presumably lessen the probability of dehiscence. The purpose of this alternating, prospective study was to evaluate the outcome of fracture of the hamulus process at the time of palate repair. The outcome assessment was accomplished by comparing the operative and perioperative complications, hearing results, and speech function in the groups that either had or did not have hamulus fracture. MATERIALS
AND

METHODS

Patients, Study Enrollment, and Demographic Variables Recorded A total of 173 consecutive patients was evaluated by one surgeon (L.J.L.) at Chang Gung Craniofacial Center prior to undergoing palatoplasty. Following evaluation, these patients all underwent palatoplasty by the same surgeon during the 2year time period spanning March 1996 through February 1998. Patients were assigned to either have or not have hamulus fracture preoperatively, on a strictly alternating basis, without regard to any patient characteristic. Demographic variables recorded for each patient included sex, birth date, description of any known associated craniofacial or systemic anomalies, and cleft type and width. With regards to cleft width, we followed our protocol (Lo et al., 1999) of intraoperatively measuring the posterior width of the cleft as a percentage of the distance between the tuberosities. If this percentage was 20% to 30%, it was a moderate-width cleft, while larger percentages and smaller ones were respectively scored as wide and narrow. The results of all the demographic variables were grouped with respect to hamulus fracture and tested for signicance using the chi-square method. Surgical Procedure Selection and Surgical Outcome Measures Hamulus fracture was carried out in an identical matter on each patient assigned to the fracture group, independent of the other specic techniques employed during the palatoplasty. The hamulus processes were fractured inwardly bilaterally, by pressure exerted from an elevator during the dissection of the velar musculature away from its abnormal insertion on the hard palate, prior to its posterior repositioning. Dissection around the greater palatine vascular pedicles was performed to release all brous tethering allowing the ap to move freely toward midline. All of the techniques used in the palatoplasties involved raising the mucoperiosteal tissues off the hard palate, differing only in the incisions made, which determine the amount of detachment of the mucoperiosteum from the surrounding tis-

sues. One of ve techniques of palatoplasty was chosen for each patient, dependent upon the opinion of the surgeon with respect to the anatomy of each cleft. The choice was coded as one of the following: (1) Bardachs two-ap technique (Bardach and Salyer, 1991), raising bilateral mucoperiosteal aps based uniquely on the greater palatine arteries; (2) Langenbeck palatoplasty, raising bilateral bipedicled mucoperiosteal aps with lateral relaxing incisions but no release of the anterior mucoperiosteum as in the two-ap technique; (3) a no-ap technique (Sommerlad, personal communication, 1998), consisting of raising the mucoperiosteal tissues medially to laterally but without creation of lateral relaxing incisions; (4) a combination of a Langenbeck technique on one hemipalate and a Bardach-type unipedicled ap on the other hemipalate; and (5) a combination of no ap on one hemipalate and a Langenbeck technique on the other hemipalate. For simplicity, these descriptions will subsequently be referred to respectively as the two-ap, Langenbeck, no-ap, Langenbeck two-ap, and Langenbeck no-ap techniques. Each technique was carried out in a highly uniform fashion each time it was selected. In all cases, the mucoperiosteal aps were raised from the midline rst and carried laterally. If it was judged that the tissues could be approximated in the midline without tension, this was done and thereby constituted the no-ap technique. As a next step to relieve tension, a lateral relaxing incision was made on one side and then on the other side if needed. If tension was still too great, then the bipedicled Langenbeck aps were converted to the unipedicled two-ap type, one at a time. No attempt was made to approximate the palatal layer to the nasal layer in any of the techniques, because it is felt that although this maneuver obliterates dead space, it also reintroduces some tension in the aps. All patients receive injection of an epinephrine 1: 200,000 solution to the palate prior to undergoing the procedure. Coagulation was almost never used. Surgicel was occasionally used to pack the lateral relaxing incisions at the end of the procedure if oozing was noted. In addition to the palatoplasty type and the occurrence of hamulus fracture, the following variables related to the operation and postoperative complications were recorded: palatoplasty date; operative time (duration of procedure); estimated blood loss; transfusions given in the intra- or perioperative periods; occurrence of postoperative bleeding requiring operative intervention; occurrence of stula (any communication between the oral and nasal cavities due to complete, full-thickness breakdown of repair and requiring subsequent operation to repair); and occurrence of oral mucosal dehiscence (a breakdown of partial thickness of repair that spontaneously closed without intervention). All of these variables, with the exception of palatoplasty date, were grouped with respect to presence or absence of hamulus fracture and then tested for signicance using chi-square analysis. Otolaryngological Outcome Measures All charts were reviewed for several otolaryngological variables. In order to assess the effects of hamulus fracture upon

508

Cleft PalateCraniofacial Journal, September 2000, Vol. 37 No. 5

Table 1 Demographic Data, Cleft Anatomy, and Associated Diagnoses Related to Presence or Absence of Hamulus Fracture
Age at Operation (mo) Age at Last Follow-Up (mo)

Number of Patients

Sex (M/F)

Cleft Type

Cleft Width

Associated Anomalies

Without hamulus fracture

76

12.6

25.7

33/43

With hamulus fracture

85

12.1

24.5

46/39

NA
bilateral cleft lip and palate; CP

NA
cleft palate; UCLP

NA

0.175

UCLP 36 BCLP 13 CP Only 27 UCLP 33 BCLP 14 CP Only 38 0.465

not applicable.

wide 25 moderate 34 narrow 17 wide 34 moderate 30 narrow 20 0.462

yes 17 no 59 yes 21 no 64 0.727

BCLP

unilateral cleft lip and palate; NA

hearing outcomes, data were collected on the subset of patients who had auditory brainstem response (ABR) testing performed. At our institution the protocol for ABR testing involves delivering a measured stimulus to each ear through earphones, recording brainstem response with electrodes placed on the scalp and skin. The stimulus starts at 100 decibels, and is decremented by 10 decibels until no response is detected by the electrodes. Normal response is dened as the interval between 0 and 30 decibels, the minimum stimulus that will elicit a brainstem response. The preoperative ABR is usually done at 8 to 11 months of age, while the postoperative ABR is done at about 18 months of age. The patient age at testing and the result for each ear were recorded for each test performed. For each patient who had at least one prepalatoplasty and at least one postpalatoplasty ABR, a comparison of the latest preoperative and earliest postoperative examinations was made. The postoperative result for each ear was subtracted from the corresponding preoperative result. The results were summed and an average result was calculated. In addition, each pre- and postresult pair was categorized as one of the following types: improved and abnormal; improved and normal; same and abnormal; same and normal; worse and abnormal; worse and normal. These results were tabulated and then divided into those from patients who received hamulus fracture and those who did not. The outcomes were tested for signicance by the chi-square method. Myringotomy tubes data were collected for patients who had a set (left and right) inserted. There were several otolaryngologists who performed the myringotomies. Date of insertion and maximum date of extrusion (dened as the date of outpatient visit when the otolaryngologist rst noted that at least one tube was either missing or no longer spanning the tympanic membrane) were recorded for each set. There was also a description of the effusion in each ear recorded by the otolaryngologist at time of insertion, coded as either serous, mucoid, or glue. The effusion descriptions were tested to see whether they were predictive of early extrusion, which was dened as extrusion less than or equal to 6 months after tube insertion. The fracture and nonfracture groups were tested to see whether this maneuver was predictive of early extrusion. Cultures were obtained in a subset of myringotomies, and the nal reports of these cultures were recorded for this subset. Since this study involved clinical interventions of several otolaryngologists, it was not possible to obtain cultures for each

set of myringotomy tubes inserted. Nor was it recorded from which ear effusion the cultures were taken (i.e., left vs. right ear). Positive culture results were tabulated with respect to organism identied. The culture results relationship to hamulus fracture was analyzed by chi-square. Finally, culture results were tabulated against description of effusion to try to discern whether the effusion description bore any statistically signicant relationship to the culture results. Speech Evaluations A subset of patients was seen by speech pathologists for initial evaluation during the study period. All of these were seen after palatoplasty by one of three speech pathologists who use similar criteria for diagnosis. For this subgroup, the perceptual speech sample was rated, with regards to velopharyngeal function as either adequate, marginal, or inadequate (Noordhoff et al., 1990). The results were related to performance of hamulus fracture and tested by chi-square. RESULTS Of the 173 patients enrolled in the study, 12 were excluded either because the charts were unavailable for review (8 patients) or were lost to clinical follow-up beyond the immediate postoperative period (4 patients). These exclusions brought the total patient count to 161, with a distribution of 85 receiving hamulus fracture, and 76 without fracture. The average age at palatoplasty was 12.1 months for the fracture group (SD 76 days) and 12.6 months for the nonfracture group (SD 67 days). The average age at last follow-up visit was 24.5 months in the fracture group and 25.7 months in the nonfracture group. These demographic data and information regarding cleft anatomy and associated diagnoses are summarized in Table 1. None of the differences between the fracture and nonfracture groups reached statistical signicance. In the subset of patients who had associated anomalies, the distribution of anomalies included: two bicoronal synostosis, one bilateral ptosis, seven congenital heart disease, two ear deformities, one constriction band, one hemifacial microsomia, six multiple congenital anomalies, three median facial dysplasia (Noordhoff et al., 1993), seven Robin sequence, two Treacher Collins, two trisomy 21, one trisomy 13, one umbilical hernia, one van der Woude, and one Wilms tumor.

Kane et al., EFFECTS OF HAMULUS FRACTURE ON PALATOPLASTY OUTCOME

509

Table 2 Operative Variables, Complications, and Myringotomy Data Related to Presence or Absence of Hamulus Fracture
Average Palatoplasty Average Time Blood Loss Oral Mucosal (min) (mL) Dehiscence

Palatoplasty Type

Oronasal Fistual

Bleeding Intervention Needed

Tube Extrusion

Culture Results

Without hamulus fracture

With hamulus fracture

two-ap 32 langenbeck two-ap 21 langenbeck 10 langenbeck no-ap 4 no ap 9 two-ap 31 langenbeck two-ap 20 langenbeck 20 langenbeck no-ap 4 no ap 10 0.569

70

8.8

yes 8 no 68

yes 4 no 72

yes 0 no 76

early 6 normal 17

growth 10 no Growth 20

66

7.2

yes 10 no 75

yes 4 no 81

yes 2 no 83

early 8 normal 18

growth 6 no Growth 22

NA

NA

0.803

0.871

0.178

0.717

0.311

NA

not applicable.

The results of the operative and perioperative variables are reported in Table 2. There was only one transfusion needed, in a single patient who required 120 mL for postoperative pallor. This patient was one of the total of two who required return to the operating room to control bleeding postoperatively, and the patient did have hamulus fracture. None of the differences between groups were statistically signicant. There were 122 sets of myringotomy tubes placed in a total of 106 patients (16 patients had two sets of tubes placed). The average number of sets of tubes per patient was 1.15. The average age at the time of insertion of the rst set of tubes for all patients was 15.4 months. Of the 122 sets of tubes, 49 had the maximum date of extrusion recorded (40%). The overall average length of time that the tubes remained in place for these patients was 9.8 months, and the average for the 26 fracture and 23 nonfracture patients respectively was 9.5 and 10.1 months. For the 122 sets of tubes (with a total of 244 possible effusions), there were 218 effusion descriptions recorded, and 26 were either not recorded by the otolaryngologist (n 24) or were recorded as not having enough uid to be deemed an effusion (n 2). Effusion descriptions are related to culture results and tube extrusion in Table 3. Once again, none of the differences between groups were statistically signicant. Since the ear from which the cultures were taken was not recorded, it was necessary to divide the culture growth data by two (thus deriving an average value over both ears), which explains why there are noninteger values in the table. The extrusion data
Table 3 Effusion Descriptions Related to Culture Results and Tube Extrusion Data
Culture Results Tube Extrusion Hamulus Fracture

Serous Mucoid Glue

2

growth 4 no growth 18.5 growth 9.5 no growth 22 growth 1.5 no growth 1.5 0.377

early 11 normal 33 early 14 normal 29 early 3 normal 6 0.842

yes 52 no 35 yes 49 no 42 yes 5 no 11 0.107

were also averaged over both ears, because it was not recorded whether one or both tubes extruded or from which ear the tube extruded. A total of 58 of the 122 sets of tubes (48%) had cultures taken at the time of tube insertion. Of these, 42/58 (72%) had no growth as the nal culture result, and 16/58 (28%) had an organism identied. Two patients had two organisms grow from a single culture. The distribution of organisms identied was: two Branhamella catarrhalis; one Clostridium perfringens; four coagulase-negative Staphylococcus species; one Escherichia coli; one Haemophilus inuenzae; two Neisseria species (one unidentied, one avescens); one Proteus mirabilis; one Staphylococcus aureus; four Streptococcus pneumoniae; and one Veillonella alcalescens. A total of 60 patients had both preoperative and postoperative ABR testing performed (37% of those enrolled in study). Of these 29/60 (48%) had hamulus fracture, and 31/60 (52%) had no hamulus fracture at time of palatoplasty. The average postoperative result (when data from both ears are combined) was an improvement of 12.1 decibels in the fracture group and 14.5 decibels in the nonfracture group. The postoperative test was performed at an average of 6.8 and 6.5 months for the fracture and nonfracture groups, respectively. The data for the ABR testing are summarized in Table 4. There were no signicant differences in outcomes between the fracture and nonfracture populations, both utilizing chi-square and Mann Whitney U testing. Speech evaluations were available for 47 of the 161 patients enrolled in the study (29%). The average age at time of speech evaluation for both the hamulus fracture and nonfracture groups, respectively, was 2.7 years. Table 5 contains the speech evaluation results. There was again no signicant difference in outcomes between the fracture and nonfracture populations. DISCUSSION The operative maneuver of hamulus fracture is credited as having been introduced by Billroth (1889), and its perfor-

510

Cleft PalateCraniofacial Journal, September 2000, Vol. 37 No. 5

Table 4 Effusion Descriptions Related to Culture Results and Tube Extrusion Data
Average Postoperative Results (decibels improved) Average Months Postop That ABR Preformed

Postoperative Result Description

Without hamulus fracture

14.5

6.8

With hamulus fracture

12.1

6.5

NA
not applicable.

NA

improved: abnormal 17, normal 24 worse: abnormal 10, normal 2 same: abnormal 2, normal 7 improved: abnormal 18, normal 19 worse: abnormal 8, normal 1 same: abnormal 5, normal 7 0.803

NA

mance has been advocated as a means of reducing tension at time of closure of the soft palate. Its adoption into the armamentarium of cleft surgical maneuvers has been based upon a presumed improvement in results due to this reduction of tension. In turn, this reduction in tension would theoretically produce an advantage in terms of outcome. Namely, there might be fewer problems with wound healing and subsequent facial growth. We are unaware of any objective studies that quantitate this decrease in tension, nor are we aware of any studies that prove that the maneuver produces the intended theoretical result. Furthermore, the effect of the maneuver upon speech has not been studied. The effects on middle ear disease of fracture of the pterygoid hamulus during palatoplasty have been previously prospectively evaluated in a well-designed study by Noone et al. (1973). In their study, each patient served as his own control, because each patient underwent only unilateral hamulus fracture, which was randomly assigned. Eighty-nine patients were evaluated with audiograms and clinical examinations over a 3year follow-up period. It is not explicitly stated whether one or more surgeons performed the operations, yet it is implied that several surgeons performed the palatoplasties. The scope of the study did not include surgical outcomes related to the palate itself or the impact upon speech. No differences were noted in outcomes of middle ear disease, and it was concluded that . . . fracture of the pterygoid hamulus, and, thus, disturbance of the tensor veli palatini tendon during soft palate closure does not signicantly alter the state of middle ear disease. Furthermore, it was concluded that . . . the elimination of hamulotomy from soft palate closure because of fear of middle ear disease would not be justied. In essence, the authors concluded there was no contraindication to the fracture of the hamulus during palatoplasty with respect to middle ear
Table 5 Preliminary Speech Results
Average Evaluation Age (years)

Speech Result

Without hamulus fracture

2.7 (SD 0.31) 2.7 (SD 0.67) NA

With hamulus fracture

adequate 20 marginal 3 inadequate 1 adequate 18 marginal 3 inadequate 2 0.812

NA

not applicable.

disease, yet it did not prove that the maneuver was justied on the basis of its intended result. Particular surgical maneuvers are utilized, depending upon preference and belief in their utility, and are often passed from generation to generation of surgeons. These beliefs may be derived from the opinions of those who trained the surgeon. In a questionnaire described by Noone et al. (1973), 42 of 101 cleft surgeons reported they routinely fractured the hamulus during palatoplasty. The source of this information was cited as unpublished data provided by Dr. Stool from the Cleft Palate Clinic at the Childrens Hospital of Philadelphia. Unfortunately, the frequency with which hamulus fracture currently is performed is not known. Such information would be valuable framing information for the interpretation of this analysis. The results of this study corroborate those of Noone et al., with respect to middle ear disease. There was no advantage in the outcomes of ABRs in the fractured vs. nonfractured groups. Complete ABR records were available for 37% of all enrollees, which provided comparable sample sizes of approximately 30 patients each. It was felt that ABR provides a more objective and reliable result than does audiometry, which was used in the above-cited study, and would be unfeasible in this age group. Note that the ABR results reect the outcomes at an average postoperative age of about 6.5 months. While these results prove no difference between the groups at 6 months, there is theoretically nothing to prevent signicant differences from emerging with longer follow-up. Still, it was felt that if differences were going to emerge from an operative maneuver such as hamulus fracture, they would be likely to have done so in the early postoperative period. This point might, of course, be argued. One can infer that if hamulus fracture confers an advantage in terms of reduced tension at time of palate closure, then such an advantage might be measurable by observing a reduction in wound complications. Yet no such diminution was observed in terms of the rates of supercial oral mucosal dehiscence or in stula occurrence. In fact, there were no advantages in any of the measured operative or postoperative variables between the two groups. Since we were unable to measure a benet of hamulus fracture, we can not recommend its performance on the basis of its historical intent or theoretical rationale. A statistical caveat needs to be added in the interpretation of some of the results of the perioperative complications. Since the overall rate of oronasal stula and bleeding intervention is so

Kane et al., EFFECTS OF HAMULUS FRACTURE ON PALATOPLASTY OUTCOME

511

low, even our fairly large sample size lacks sufcient statistical power to condently reveal these differences. Consequently, we could erroneously be accepting the null hypothesis here, committing a type II error. The myringotomy tube data analysis also failed to show any differences between the fracture and nonfracture groups. Performance or absence of the maneuver had no predictive value in determining the extrusion pattern of the tubes, the culture results of the effusions, or the description of the effusions. The fact that the average tube insertion age was approximately 3 months following the average palatoplasty date may explain this observation. Yet again, it is possible that more sampling or a different temporal pattern of effusion sampling would reveal a difference between the groups. Although it was unrelated to the main subject of this paper, it was decided to test the relationship between the effusion description and culture result and between effusion description and tube extrusion. The results showed that the otolaryngologists perception of the effusion bore no statistically signicant relationship to culture results or in predicting the duration of utility of a set of tubes. Preliminary speech results also showed no differences between the fracture and nonfracture groups. This result is of limited value because only 29% of the patients made it to an age at which speech evaluation could be done during the duration of the study. It will be interesting to follow up these patients over time, because this result is necessarily weak due to the short follow-up period and young average age at time of evaluation (average age at speech follow-up was 27 months), and this is our intent. The speech results overall were acceptable, with 81% assessed as adequate, 13% as marginal, and 6% as inadequate. It is difcult to compare these overall results with those of others, because the standards and variables by which such comparisons are made vary considerably between individuals, yet it should be noted that all of these patients did have the velar musculature posteriorly repositioned and sutured in the midline. In general, the subjects of follow-up time and missing data need to be addressed. In this study, the average patient was followed up for about a year following palatoplasty. In interpreting these results, it must be remembered that the lack of difference in outcomes between the fracture and nonfracture at a follow-up interval of 1 year does not necessarily preclude the possibility of differences emerging with more lengthy observation. Still, we believe it is reasonable to conclude that there would be no changes in the perioperative variables with longer follow-up. With regards to missing data points, in the setting in which these operations were performed, the protocol was followed in every patient who had speech and ear, nose, and throat (ENT) follow-up at our institution. Unfortunately, in our setting, it is not possible to assure that every patient has speech or ENT follow-up at our institution, and, therefore, we were not able to obtain data for these patients through the exercise of our protocol. The missing data are a reection of the percentage of patients who have not been followed up at our institution.

SUMMARY

AND

CONCLUSION

In this randomized and prospective study, there were no differences in perioperative morbidity, hearing results, or very preliminary speech results between patients who received hamulus fracture and those who did not. We are therefore unable to advocate the performance of hamulus fracture as an operative maneuver during the performance of primary palatoplasty. The historical rationale and theoretical advantage of this maneuver have not been demonstrated in this study, nor have any detrimental effects of the maneuver been measured.
Acknowledgments. We acknowledge and are grateful for the support given to us by the speech pathologists afliated with our Craniofacial Center: Claudia Yun, Ruby Wang, and Jorie Wu as well as the technical assistance provided by Pei-Yu Chen. This work was supported by a grant from the National Science Council, R.O.C., NSC 87-2314-B-182A-074.

REFERENCES
Bardach J, Salyer KE. Cleft palate repair. In: Bardach J, Salyer KE, eds. Surgical Techniques in Cleft Lip and Palate. St. Louis: Mosby Year Book; 1991:224273. Barsoumian R, Kuehn DP, Moon JB, Canady JW. An anatomic study of the tensor veli palatini and dilatator tubae in relation to eustachian tube and velar function. Cleft Palate Craniofac J. 1998;35:101110. Billroth T. Ueber Uranoplastik (Krankenvorstellung). Wien Klin-Wchnschr. 1889;2:241. Bluestone CD, Beery QC, Cantekin EI, Paradise JL. Eustachian tube ventilatory function in relation to cleft palate. Ann Otol Rhinol Laryngol. 1975;84:333 338. Chaudhuri PK, Bowen-Jones E. An otorhinological study of children with cleft palates. J Laryngol Otol. 1978;92:2940. Fria TJ, Paradise JL, Sabo DL, Elster BA. Conductive hearing loss in infants and young children with cleft palate. J Pediatr. 1987;111:8487. Gopalakrishna A, Goleria KS, Raje A. Middle ear function in cleft palate. Br J Plast Surg. 1984;37:558565. Lo LJ, Huang CS, Chen YR, Noordhoff MS. Palatoalveolar outcome at 18 months following simultaneous primary cleft lip and posterior palatoplasty. Ann Plast Surg. 1999;42:581588. Noone RB, Randall P, Stool SE, Hamilton R, Winchester RA. The effect on middle ear disease of fracture of the pterygoid hamulus during palatoplasty. Cleft Palate J. 1973;10:2333. Noordhoff MS, Huang CS, Lo LJ. Median facial dysplasia in unilateral and bilateral cleft lip and palate: a subgroup of median cerebrofacial malformations. Plast Reconstr Surg. 1993;91:9961005. Noordhoff MS, Huang CS, Wu J. Multidisciplinary management of cleft lip and palate in Taiwan. In: Bardach J, Morris HL, eds. Multidisciplinary Management of Cleft Lip and Palate. Philadelphia: WB Saunders; 1990:1826. Paradise JL. Middle ear problems associated with cleft palate. An internationally-oriented review. Cleft Palate J. 1975;12:1722. Ross MA. Functional anatomy of the tensor palati. Its relevance in cleft palate surgery. Arch Otolaryngol. 1971;93:13. Shibahara Y, Sando I. Histopathologic study of eustachian tube in cleft palate patients. Ann Otol Rhinol Laryngol. 1988;97:403408. Spauwen PH, Hillen B, Lommen E, Otten E. Three-dimensional computer reconstruction of the eustachian tube and paratubal muscles. Cleft Palate Craniofac J. 1991;28:217219. Stool SE, Randall P. Unexpected ear disease in infants with cleft palate. Cleft Palate J. 1967;4:99103. Yules RB. Hearing in cleft palate patients. Arch Otolaryngol. 1970;91:319 323.