Correspondence
Alternative Techniques of Hysterectomy
N Engl J Med 1997; 336:290-293January 23, 1997
- Article
To the Editor:
With regard to the articles on alternative techniques of hysterectomy by Dorsey et al. and Weber and Lee and the accompanying editorial by Stovall and Summitt (Aug. 15 issue),1-3 the terms “laparoscopic hysterectomy” and “laparoscopically assisted vaginal hysterectomy” require clarification. Laparoscopic hysterectomy entails the complete separation of the uterus from its vascular, vaginal, and connective-tissue attachments, which is accomplished entirely by means of laparoscopic manipulation. Laparoscopically assisted vaginal hysterectomy may be no more than a laparoscopic visualization of the pelvic viscera followed by a standard vaginal hysterectomy. In between these two procedures are those that involve stapling, burning, or ligating the various uterine attachments, including separation of the bladder from the lower uterine segment. Although all these procedures fall under the general heading of laparoscopically assisted vaginal hysterectomy, they vary enormously in terms of operating time, cost of disposable surgical materials, and required operative skill.
Several trends are coming into clearer focus. When the cost of disposable equipment is charged to the physician, laparoscopically assisted vaginal hysterectomy will disappear. With the newly recognized subspecialty of pelvic reconstructive surgery and urogynecology, hysterectomies will be performed in greater numbers by fewer and more experienced gynecologists. With more experience in performing vaginal hysterectomies, gynecologists become less dependent on the laparoscope to remove the uterus. The differences in morbidity and length of convalescence between abdominal and vaginal hysterectomy are diminishing. Patients now leave the hospital two to three days after an abdominal hysterectomy and return to work weeks earlier than they used to. In addition, the clear demonstration of the cardioprotective effect of postmenopausal estrogen therapy4 will contribute to an increased use of hormone-replacement therapy for women. The benefit of such therapy will result in an increased number of hysterectomies in women with myomas, family histories of endometrial cancer, or an intolerance of progestogens.
4 ReferencesRobert F. Porges, M.D.
New York University School of Medicine, New York, NY 100161
Dorsey JH ,Holtz PM ,Griffiths RI ,McGrath MM ,Steinberg EP . Costs and charges associated with three alternative techniques of hysterectomy. N Engl J Med 1996;335:476-482
Full Text | Web of Science | Medline2
Weber AM ,Lee J-C . Use of alternative techniques of hysterectomy in Ohio, 1988-1994. N Engl J Med 1996;335:483-489
Full Text | Web of Science | Medline3
Stovall TG ,Summitt RL Jr . Laparoscopic hysterectomy -- is there a benefit? N Engl J Med 1996;335:512-513
Full Text | Web of Science | Medline4
Grodstein F ,Stampfer MJ ,Manson JE , et al. Postmenopausal estrogen and progestin use and the risk of cardiovascular disease. N Engl J Med 1996;335:453-461
Full Text | Web of Science | Medline
To the Editor:
Dorsey et al. compare the costs of three techniques of hysterectomy, using data from 1049 procedures performed by 96 surgeons. In their discussion, the authors acknowledge that they did not control for the effect of the surgeon on total costs, operating-room time, and length of hospitalization. However, the surgeon would be expected to affect these variables, at least to some extent. Furthermore, it is not unlikely that particular surgeons tend to select one type of hysterectomy over another. Thus, the differences in costs, attributed by the authors to factors intrinsic to the surgical procedures, could readily be confounded by differences among surgeons in the speed of their work, their technical skill, or the post-operative care they provide.
The authors analyzed the effects of the type of surgical procedure and covariates on the costs of treatment, using multiple regression and t-tests (equivalent to simple regression). Each surgeon's cases actually represent clusters of data. Ignoring the effects of the individual physician is equivalent to assuming that the within-cluster correlation is zero. Zeger and Liang1 have explained how ignoring the within-cluster correlation in such situations leads to a loss of precision in estimating the variance of the regression parameters. In the case of a t-test for comparing group means, if the within-cluster correlation were other than zero, the computed t-statistic would be invalid. Thus, the authors' inferences regarding differences among the types of hysterectomy may be invalid.
Other available regression techniques would have allowed the authors to account for the effects of individual surgeons. A marginal regression model, such as the generalized estimating equation, could have been used.1,2 Such an approach would have made sense in this study, because the authors were not interested in estimating the specific effects of individual physicians on the cost of hysterectomy. Using the generalized estimating equation, they could have estimated the marginal effects, for the group of surgeons as a whole, of covariates such as the type of hysterectomy on costs, the duration of the procedure, and the length of the hospital stay, while accounting for the within-cluster correlation.
2 ReferencesJohn W. Robinson, M.D., Ph.D.
CMG Health, Owings Mills, MD 211171
Zeger SL ,Liang KY . An overview of methods for the analysis of longitudinal data. Stat Med 1992;11:1825-1839
CrossRef | Web of Science | Medline2
Zeger SL ,Liang KY . Longitudinal data analysis for discrete and continuous outcomes. Biometrics 1986;42:121-130
CrossRef | Web of Science | Medline
To the Editor:
Dorsey et al. raise some doubts about the advantages of laparoscopically assisted vaginal hysterectomy, because of the higher costs of the procedure as compared with other types of hysterectomy. These higher costs were confirmed by Weber and Lee. However, we believe that the two groups of authors are mistaken in basing their analyses on a series of 1049 hysterectomies performed by a large number of surgeons over two years (in the case of Dorsey et al.) and retrospectively computerized data from 180 hospitals (in the case of Weber and Lee).
We have prospectively analyzed a consecutive series of 508 hysterectomies performed for benign uterine diseases by three surgeons from May 1994 to May 1996 (Table 1Table 1
Characteristics of 508 Hysterectomies, According to the Surgical Procedure Used.). Many of the operations were laparoscopic subtotal hysterectomies. After coagulation of the uterine artery, section of the upper part of the cervix was carried out. The corpus of the uterus was removed by morcellation with the Steiner device.2 This technique can be used if hysterectomy is indicated1 and if there is no risk of remaining cervical cancer.There was a significantly lower rate of abdominal hysterectomy in our study (12 percent) than in the study by Dorsey et al. (54 percent). Their high rate is surprising but can be explained by the fact that in their series, the 1049 hysterectomies were performed by 96 surgeons (with a mean of 5 hysterectomies a year per surgeon). This fact could also explain the surprisingly high rate of conversion from a laparoscopic procedure to laparotomy (12 percent). In our series, although the uterine weights in the subgroups of patients were similar to those in the study by Dorsey et al., we never converted to laparotomy, even in cases of operative bleeding or bladder injury (in three cases).3 The high rate of conversion in their study may be explained by the relative inexperience of some of the surgeons.
Another major concern is the comparison of costs. We evaluated the costs per operation, including hospital costs, the costs of medical and surgical supplies, and physicians' costs. Dorsey et al. found that laparoscopic procedures had the highest costs. However, we found that the costs of the two laparoscopic procedures were substantially lower than the costs of vaginal and abdominal hysterectomies. The difference may be due to three factors: disposable instruments were not used in our series; there was no significant difference in the operating-room time required for abdominal hysterectomies (112 minutes) and laparoscopic procedures (laparoscopic subtotal hysterectomy, 118 minutes; laparoscopically assisted vaginal hysterectomy, 129 minutes); and the hospital stay was shorter after laparoscopic procedures, especially laparoscopic subtotal hysterectomy, than after laparotomy (Table 1).
Almost all our patients who underwent laparoscopic subtotal hysterectomy were theoretically able to leave the hospital the first day after surgery. Nevertheless, many patients preferred to stay 2 or 3 days, knowing that the Belgian insurance system allows reimbursement for up to 10 days of hospitalization. In our series, the average length of stay after surgery was 2.9 days for laparoscopic subtotal hysterectomy, 4.5 days for laparoscopically assisted vaginal hysterectomy, 5.9 days for vaginal hysterectomy, and 7.1 days for abdominal hysterectomy. (The length of stay depended, for the most part, on the age of the patient.)
Our findings are the complete opposite of those of Dorsey et al.
3 ReferencesMichelle Nisolle, M.D., Ph.D.
Jacques Donnez, M.D., Ph.D.
Catholic University of Louvain, B-1200 Brussels, Belgium1
Donnez J ,Nisolle M . LASH: laparoscopic supracervical hysterectomy. J Gynecol Surg 1993;9:91-94
CrossRef | Medline2
Steiner RA ,Wight A ,Tadir Y ,Haller U . Electrical cutting device for laparoscopic removal of tissue from the abdominal cavity. Obstet Gynecol 1993;81:471-474
Web of Science | Medline3
Donnez J, Nisolle M, Anaf V. Laparoscopy-assisted vaginal hysterectomy and laparoscopic hysterectomy in benign diseases. In: Donnez J, Nisolle M, eds. An atlas of laser operative laparoscopy and hysteroscopy. New York: Parthenon Publishing, 1994:203-11.
To the Editor:
In their editorial, Stovall and Summitt erroneously attribute to my colleagues and me the suggestion that “laparoscopically assisted vaginal hysterectomy should replace the traditional vaginal approach.” In neither the report they cite1 nor a more recent report2 did we suggest that laparoscopically assisted vaginal hysterectomy is a substitute for vaginal hysterectomy. Laparoscopically assisted vaginal hysterectomy may be a substitute for abdominal hysterectomy if the anatomical characteristics or the condition of the adnexa make the success of operation by the vaginal route uncertain.
2 ReferencesNewton G. Osborne, M.D., Ph.D.
Howard University College of Medicine, Washington, DC 200601
Padial JG ,Sotolongo J ,Casey MJ ,Johnson C ,Osborne NG . Laparoscopy-assisted vaginal hysterectomy: report of seventy-five consecutive cases. J Gynecol Surg 1992;8:81-85
CrossRef | Medline2
Garcia-Padial J ,Osborne N ,Sotolongo J ,Ferrer N . Laparoscopy-assisted vaginal hysterectomy compared with abdominal hysterectomy. J Natl Med Assoc 1995;87:288-290
Web of Science | Medline
To the Editor:
It is debatable whether anything of lasting value can yet be said about the cost of laparoscopic surgery, because most of the techniques are still inchoate, and a wholesale shift from conventional to laparoscopic surgery would so fundamentally change the fabric of hospital care as to defy meaningful mensuration. In addition, Dorsey et al. mislead the unwary about the cost to society of laparoscopic techniques by improperly emphasizing and misstating the record on hospital charges (since laparoscopic hysterectomies cost less than abdominal hysterectomies when the former are performed by experienced surgeons without the use of expensive staplers1) and by failing to address the effect of laparoscopic surgery on recovery time and indirect costs.
In their editorial, Stovall and Summitt further distort the record by claiming that most hysterectomies can be performed vaginally (whereas history teaches us that they cannot) and that laparoscopic hysterectomies have no advantages over, and cost more than, vaginal hysterectomies. The authors rely on their own randomized studies.2,3 The first of these studies was not informative about comparative morbidity, because postoperative pain and recovery time were not documented, and it had a very large type II error (fewer than 30 patients per group were studied). In addition, the study's hypothesis was simply not credible, because laparoscopic and vaginal hysterectomies were compared only in highly selected women (those who were candidates for discharge on the day of surgery), in whom morbidity associated with any kind of surgery could be expected to be so low that an enormously large study would be required to detect any differences.2 Nonetheless, the serious complications (in 2 of 27 patients) after vaginal hysterectomy included a vesicovaginal fistula and a pelvic hematoma, and the mean blood loss was significantly higher after this operation. Although the charges for laparoscopic hysterectomy were higher, endoscopic staplers were used in all cases. Subsequently, we learned that expensive pretreatment with gonadotropin-releasing–hormone agonists was required before these authors could vaginally remove fibroid uteruses 14 to 18 weeks in size, and even then, they were successful in only 80 percent of the cases,3 whereas with laparoscopic assistance, uteruses of this size or larger can be removed without treatment with gonadotropin-releasing–hormone agonists virtually 100 percent of the time.4
4 ReferencesNicholas Kadar, M.D.
New Margaret Hague Women's Health Institute, Secaucus, NJ 07094Harry Reich, M.D.
Columbia University College of Physicians and Surgeons, New York, NY 10032C.Y. Liu, M.D.
Chattanooga Women's Laser Center, Chattanooga, TN 37421Ray Garry, M.D.
, London WC1N 2BL, United Kingdom1
Munro MG ,Deprest J . Laparoscopic hysterectomy: does it work? A bicontinental review of the literature and clinical commentary. Clin Obstet Gynecol 1995;38:401-425
Web of Science | Medline2
Summit RL Jr ,Stovall TG ,Lipscomb GH ,Ling FW . Randomized comparison of laparoscopy-assisted vaginal hysterectomy with standard vaginal hysterectomy in an outpatient setting. Obstet Gynecol 1992;80:895-901
Web of Science | Medline3
Stovall TG ,Summit RL Jr ,Washburn SA ,Ling FW . Gonadotropin-releasing hormone agonist use before hysterectomy. Am J Obstet Gynecol 1994;170:1744-1751
Web of Science | Medline4
Kadar N. Extraperitoneal laparoscopic hysterectomy for the large uterus. Gynaecol Endosc (in press).
Author/Editor Response
The authors reply:
To the Editor: Dr. Porges suggests that if surgeons, rather than hospitals or insurers, were to bear the cost of disposable equipment, laparoscopy-assisted vaginal hysterectomy would “disappear.” Although such a change might decrease the use of disposable supplies, our results suggest that it need not decrease the use of laparoscopy-assisted vaginal hysterectomy.
Dr. Robinson suggests that the use of more sophisticated regression techniques, such as the generalized-estimating-equation approach, which could have controlled for differences in surgeons' technical skills, might have demonstrated that, in fact, there were no differences in cost among the techniques. Our failure to control for surgeon-specific effects could have resulted in a bias. It is possible, for example, that the surgeons who used disposable equipment for all steps of the laparoscopic portion of laparoscopy-assisted vaginal hysterectomy were more (or less) skilled than the surgeons who did not use disposable supplies at all or used them for only some steps. An ideal study would examine the outcomes of different surgical techniques performed by surgeons with equal skill. We looked at the outcomes of surgery as performed in real life (an “effectiveness” rather than an “efficacy” study). Because surgeons who use disposable supplies tend to use them with all their patients, neither the generalized estimating equation nor any other statistical approach would be helpful in identifying this type of bias.
Dr. Robinson also points out that we analyzed data on multiple patients operated on by individual surgeons and that these repeated measures are likely to be correlated. As a result, we may have overestimated or underestimated the precision of our statistical estimates. Although the generalized estimating equation would have provided more precise P values, it is unlikely that our conclusions would have changed, since our P values tended to be so small.
Drs. Nisolle and Donnez suggest that their results cast doubt on the validity of our findings. However, nearly half their cases involved subtotal rather than total hysterectomy. In addition, they did not control for differences in secondary procedures that were performed in conjunction with the hysterectomy. Differences in such secondary procedures may account for differences not only in the rate of conversion from a laparoscopic procedure to a laparotomy but also in the length of stay and cost.
We agree with Dr. Osborne that laparoscopy-assisted vaginal hysterectomy may be usefully substituted for total abdominal hysterectomy in some cases, but laparoscopy-assisted vaginal hysterectomy should not be substituted for vaginal hysterectomy.
Dr. Kadar et al. appear to have misread our report. We assessed charges, as well as medical-center costs (not costs to society), for the different hysterectomy techniques. We found that laparoscopy-assisted vaginal hysterectomy was not substantially more expensive than total abdominal hysterectomy when nondisposable supplies were used. Finally, as we stated in our article, because of the relevance of potential differences in patients' quality of life, recovery time, and ability to return to work after various types of hysterectomy, we are studying the outcomes of each type after discharge, as well as in the hospital.
James H. Dorsey, M.D.
Patrice M. Holtz, R.N., M.S.
Greater Baltimore Medical Center, Baltimore, MD 21204Earl P. Steinberg, M.D., M.P.P.
Corning HTA, Washington, DC 20005-3934Author/Editor Response
We appreciate the comments of Drs. Nisolle and Donnez. One strength of an analysis such as ours is the different perspective it provides by using data from many hospitals over a period of several years, instead of using data from a single institution with a few surgeons over a short period, as Drs. Nisolle and Donnez have done. Our analysis applies to surgical experience on a larger scale, representing a wider range of persons practicing surgical gynecology.
Anne M. Weber, M.D.
Jar-chi Lee, M.S.
Cleveland Clinic Foundation, Cleveland, OH 44195- Citing Articles (1)
Citing Articles
1
Bora Cengiz, Lutfi Cem Demirel, Fulya Dokmeci, Mete Güngör, Akin Canga, Sevim Dincer Cengiz. (2002) Bilateral Salpingo-Oophorectomy During Vaginal Hysterectomy in Cases with Nonprolapsed Uterus: Role of Laparoscopy in a Residency Training Program Without Much Vaginal Salpingo-Oophorectomy Experience. Journal of Gynecologic Surgery 18:3, 87-93
CrossRef
