294
Letters
July 1992 Am J Obstet Gynecol
tent and severity of the disease should be made before laser therapy and that every mode available for evaluation and appraisal should be used. Multiple selected biopsies are most important to rule out the possibility of underlying invasive disease. However, we believe that once invasive disease has been ruled out a laser ablation of the affected area and its immediate surrounding tissue is the best way to eliminate preinvasive disease and give the patient the best result with the least degree of morbidity and complications and with improved subsequent sexual function. Some authors' who have dealt with this problem cite the difficulty of the technical ability to eradicate adequately those pits and deep scars that may occur in the vault of the vagina and in the corner recesses. We believe our technique allows us to apply the laser therapy accurately and adequately to accomplish the result. The use of appropriately placed wire sutures allows better visualization and accessibility of the laser beam to the tissue. In addition, the use of submucosal infiltration allowed us to accomplish our published results. Since the publication of our article our continued successful results have convinced us of the appropriateness of the procedure with the guidelines suggested. Furthermore, we are sorry that Hoffman et al. did not find it convenient to include our article in their list of references. Alfred I. Sherman, MD Department of Obstetrics and Gynecology, Sinai Hospital, 6767 West Outer Dr., Detroit, MI 48235
REFERENCE 1. Sherman AI. Laser therapy for vaginal intraepithelial neo-
plasia after hysterectomy. J Reprod Med 1990;35:941-4.
Reply To the Editors: I indeed overlooked Sherman's article,
and I apologize to him for this. I have now reviewed his article and find that it contains a large series of patients who were successfully treated by laser vaporization.' The successful treatment in his study was attributed to patient selection, infiltration of fluid into the submucosa, and the use of wire sutures for exposing scarred and recessed areas. We have also used submucosal infiltration in some patients and have found it helpful for hemostasis, but we have not used it in the manner described by Sherman. We and others have also used skin hooks in a manner probably similar to the wire sutures to gain exposure to recessed areas. We find the techniques described by Sherman to be interesting and will certainly consider using them in the future. Our findings and conclusions on the basis of our experience with grade 3 vaginal intraepithelial neoplasia are quite different from those of Sherman's, and there may be reasons for these differences besides technique. Although all of our patients had grade 3 intraepithelial neoplasia, it is unclear whether this was the case in Sherman's paper. Our study was much smaller than Sherman's and it would not be appropriate to make a more in-depth comparison between the two studies. We, however, are particularly concerned about
the use of laser vaporization for grade 3 vaginal intraepithelial neoplasia diagnosed at or near a vaginal cuff scar, particularly when there is marked distortion around this area or when the previous hysterectomy had been done for squamous cell neoplasia. In our limited experience with such patients we have seen a high rate both of recurrence and of occult invasive carcinoma. '"3 Mitchel S. Hoffman, MD Division of Gynecologic Oncology, Moffit Cancer Center, University of South Florida, 12902 Magnolia Dr., Tampa, FL 33612-9497
REFERENCES 1. Sherman AI. Laser therapy for vaginal intraepithelial neoplasia after hysterectomy. J Reprod Med 1990;35:941-4. 2. Hoffman MS, Roberts WS, LaPolla JP, Sterghos S Jr, Cavanagh D. Neoplasia in vaginal cuff inclusion cysts after hysterectomy. J Reprod Med 1989;34:412-4. 3. Hoffman MS, DeCesare SL, Roberts WS, FioricaJV, Finan MA, Cavanagh D. Upper vaginectomy for in situ and occult, superficial invasive carcinoma of the vagina. AM J OBSTET GVNECOL 1992; 166:30-3.
Initiation of antepartum testing in the hypertensive gravid woman To the Editors: We read with interest the article by Pircon
et al. (Pircon RA, Lagrew DC, Towers CV, Dorchester WL, Gocke SE, Freeman RK. Antepartum testing in the hypertensive patient: when to begin. AM J OBSTET GYNECOL 1991;164:1563-70). We are concerned that their conclusions are not substantiated by the data as presented. The stated purpose of their investigation was to determine the gestational age at which antepartum testing should begin in various subgroups of hypertensive gravida women. However, the data do not tell the reader how many or with what frequency antepartum tests were done at various gestational ages. We are concerned that the study analyzed the data as the percentage of abnormal antepartum tests per number of patients. To determine the sensitivity and specificity of any test, the results must be analyzed as the percentage of abnormal tests per number of tests performed at a given gestational age. Only in this way can meaningful conclusions be drawn regarding the differential use of these tests in various subgroups of patients with hypertension. For example, if diabetic and nondiabetic patients with hypertension are equally at risk for abnormal test results but the patients with diabetes are tested earlier in pregnancy and more often, it is to be expected that the patients with diabetes would have a higher percentage of abnormal test results per patient. However, the patients with diabetes would have the same percentage of abnormal test results per number of tests performed. In addition, we were unable to delineate the test re-
sults by group and by subsequent management. Are the 53 patients with positive contraction stress tests included in the 119 with "abnormal test results"? Were the abnormal test results separated into the categories of abnormal contraction stress test, abnormal nonstress
Volume 167 Number 1
test, and abnormal biophysical profile? The data also fail to indicate how often and under what circumstances an abnormal test result was ignored and the pregnancy allowed to continue for an extended period. The statement of Pircon et al. that antepartum testing in diabetic patients with hypertension should be started at viability is apparently based on the two late secondtrimester stillbirths that occurred during the study period. Pircon et al. acknowledge that both fetuses had true knots in the umbilical cords; these knots could have had a role in their deaths. Are there data demonstrating that diabetic patients with hypertension are at increased risk of umbilical cord knotting resulting in stillbirths? The question of when to begin fetal testing in patients with hypertension is extremely important. A retrospective analysis could be quite informative. We believe that the data presented by Pircon et al. should be reanalyzed, taking into account the number of tests performed at any given gestational age, with a more complete data presentation. Then justifiable conclusions could be drawn as to when antepartum testing should be initiated in the patient with hypertension. Avick C. Mitra, MD, and William N.? Herbert, MD Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of North Carolina School of Medicine, CB# 7570, Chapel Hill, NC 27599-7570
Reply To the Editors: We thank Mitra and Herbert for their interest in our study of antepartum testing in patients with hypertension. The answers to their questions can be found in the text of our original publication; however, we are happy to review the results again. Our objectives were to evaluate when fetal compromise occurs in this patient population and to make recommendations as to when to initiate antepartum testing. Contrary to the suggestion of Mitra and Herbert, the data were not analyzed as a percentage of abnormal antepartum tests per number of patients. We were not interested in the age-old issue of sensitivity or specificity of a testing method in patients with hypertension; rather, we were interested in when compromise occurs. We defined compromise by several parameters, one of which was a positive contraction stress test. Test results were recorded prospectively and were retrospectively queried by first positive test, not by the number of positive tests the patients had. In evaluating the data in this fashion we found that hypertensive patients with concomitant diagnoses of diabetes demonstrated compromise at an earlier gestational age when compared with hypertensive patients without diabetes. Subsequent management was not the topic of study. When patient management is modified because of an antepartum test, there is evidence of fetal compromise. To reiterate, we were interested in when this occurred. We all agree that follow-up of an abnormal test can be managed numerous ways. Although it seems that Mitra and Herbert have expressed an interest in our management protocol, that protocol would not affect our
Letters
295
results or our conclusions in this study. Consequently we do not believe it is appropriate to review that topic here. We recommended that testing in the diabetic patient with hypertension start at viability. As listed in our original publication, this was based on two observations. First, there were two patients with late second-trimester stillbirths and concomitant diagnoses of a true knot in the umbilical cord. It is difficult to attribute the losses solely to the knots in the cords because this is a common finding at the deliveries of normal, viable term infants. We did not suggest that there is an increased risk of cord knotting in diabetic patients with hypertension. Second, a clinician would prefer to initiate testing in a patient before fetal compromise. The earliest gestational age that intervention was necessary in the diabetic patient with hypertension was 28 weeks. Therefore testing would need to start at 26 to 27 weeks or near the time of viability. Richard Pircon, MD St. joseph's Medical Center, 5000 W. Chambers St., Milwaukee, WI 53210
Roger Freeman, MD Memorial Women's Hospital, 2801 Atlantic Ave., Long Beach, CA 90801
Legal thoughts on malpractice claims: Causes and prevention
To the Editors: As a professor of consumer law married to a physician, I read with great interest Ward's article (Ward CJ. Analysis of 500 obstetric and gynecologic malpractice claims: causes and prevention 1991; 165:298-306). I believe the article would have been improved and Kirkley'S comments muted by a change in the classification method. Ward divides the cases into "defensible" and "indefensible," and Kirkley has a number of problems with this categorization. So do I, but for a different reason. There is simply no division in the law which corresponds to the defensible-indefensible dichotomy. The law divides cases into three categories, not two. In the first the plaintiff has no case and the defendant wins on summary judgment. In the second the defendant has no case and the plaintiff wins on summary judgment. In the third the issue is presented to the jury for determination of the disputed facts. Given Ward's description of his method it would appear that most, though not all, of his indefensible cases fall into the second category, where the plaintiff might win on summary judgment. The defensible cases would be in the other two categories. However, even if a case can be made for a physician'S actions, it does not mean that the physician will win. Ward's article may give physicians the unwarranted belief that if they lose a "defensible" case, the system has somehow failed them presumably because of perjured expert testimony, idiot jurors, and corrupt lawyers. This is simply not the law, just as it is not the law that a bad outcome means that there was malpractice. The legal system recognizes that there are often factual disputes on which reasonable people can differ and
Letters
July 1992 Am J Obstet Gynecol
tent and severity of the disease should be made before laser therapy and that every mode available for evaluation and appraisal should be used. Multiple selected biopsies are most important to rule out the possibility of underlying invasive disease. However, we believe that once invasive disease has been ruled out a laser ablation of the affected area and its immediate surrounding tissue is the best way to eliminate preinvasive disease and give the patient the best result with the least degree of morbidity and complications and with improved subsequent sexual function. Some authors' who have dealt with this problem cite the difficulty of the technical ability to eradicate adequately those pits and deep scars that may occur in the vault of the vagina and in the corner recesses. We believe our technique allows us to apply the laser therapy accurately and adequately to accomplish the result. The use of appropriately placed wire sutures allows better visualization and accessibility of the laser beam to the tissue. In addition, the use of submucosal infiltration allowed us to accomplish our published results. Since the publication of our article our continued successful results have convinced us of the appropriateness of the procedure with the guidelines suggested. Furthermore, we are sorry that Hoffman et al. did not find it convenient to include our article in their list of references. Alfred I. Sherman, MD Department of Obstetrics and Gynecology, Sinai Hospital, 6767 West Outer Dr., Detroit, MI 48235
REFERENCE 1. Sherman AI. Laser therapy for vaginal intraepithelial neo-
plasia after hysterectomy. J Reprod Med 1990;35:941-4.
Reply To the Editors: I indeed overlooked Sherman's article,
and I apologize to him for this. I have now reviewed his article and find that it contains a large series of patients who were successfully treated by laser vaporization.' The successful treatment in his study was attributed to patient selection, infiltration of fluid into the submucosa, and the use of wire sutures for exposing scarred and recessed areas. We have also used submucosal infiltration in some patients and have found it helpful for hemostasis, but we have not used it in the manner described by Sherman. We and others have also used skin hooks in a manner probably similar to the wire sutures to gain exposure to recessed areas. We find the techniques described by Sherman to be interesting and will certainly consider using them in the future. Our findings and conclusions on the basis of our experience with grade 3 vaginal intraepithelial neoplasia are quite different from those of Sherman's, and there may be reasons for these differences besides technique. Although all of our patients had grade 3 intraepithelial neoplasia, it is unclear whether this was the case in Sherman's paper. Our study was much smaller than Sherman's and it would not be appropriate to make a more in-depth comparison between the two studies. We, however, are particularly concerned about
the use of laser vaporization for grade 3 vaginal intraepithelial neoplasia diagnosed at or near a vaginal cuff scar, particularly when there is marked distortion around this area or when the previous hysterectomy had been done for squamous cell neoplasia. In our limited experience with such patients we have seen a high rate both of recurrence and of occult invasive carcinoma. '"3 Mitchel S. Hoffman, MD Division of Gynecologic Oncology, Moffit Cancer Center, University of South Florida, 12902 Magnolia Dr., Tampa, FL 33612-9497
REFERENCES 1. Sherman AI. Laser therapy for vaginal intraepithelial neoplasia after hysterectomy. J Reprod Med 1990;35:941-4. 2. Hoffman MS, Roberts WS, LaPolla JP, Sterghos S Jr, Cavanagh D. Neoplasia in vaginal cuff inclusion cysts after hysterectomy. J Reprod Med 1989;34:412-4. 3. Hoffman MS, DeCesare SL, Roberts WS, FioricaJV, Finan MA, Cavanagh D. Upper vaginectomy for in situ and occult, superficial invasive carcinoma of the vagina. AM J OBSTET GVNECOL 1992; 166:30-3.
Initiation of antepartum testing in the hypertensive gravid woman To the Editors: We read with interest the article by Pircon
et al. (Pircon RA, Lagrew DC, Towers CV, Dorchester WL, Gocke SE, Freeman RK. Antepartum testing in the hypertensive patient: when to begin. AM J OBSTET GYNECOL 1991;164:1563-70). We are concerned that their conclusions are not substantiated by the data as presented. The stated purpose of their investigation was to determine the gestational age at which antepartum testing should begin in various subgroups of hypertensive gravida women. However, the data do not tell the reader how many or with what frequency antepartum tests were done at various gestational ages. We are concerned that the study analyzed the data as the percentage of abnormal antepartum tests per number of patients. To determine the sensitivity and specificity of any test, the results must be analyzed as the percentage of abnormal tests per number of tests performed at a given gestational age. Only in this way can meaningful conclusions be drawn regarding the differential use of these tests in various subgroups of patients with hypertension. For example, if diabetic and nondiabetic patients with hypertension are equally at risk for abnormal test results but the patients with diabetes are tested earlier in pregnancy and more often, it is to be expected that the patients with diabetes would have a higher percentage of abnormal test results per patient. However, the patients with diabetes would have the same percentage of abnormal test results per number of tests performed. In addition, we were unable to delineate the test re-
sults by group and by subsequent management. Are the 53 patients with positive contraction stress tests included in the 119 with "abnormal test results"? Were the abnormal test results separated into the categories of abnormal contraction stress test, abnormal nonstress
Volume 167 Number 1
test, and abnormal biophysical profile? The data also fail to indicate how often and under what circumstances an abnormal test result was ignored and the pregnancy allowed to continue for an extended period. The statement of Pircon et al. that antepartum testing in diabetic patients with hypertension should be started at viability is apparently based on the two late secondtrimester stillbirths that occurred during the study period. Pircon et al. acknowledge that both fetuses had true knots in the umbilical cords; these knots could have had a role in their deaths. Are there data demonstrating that diabetic patients with hypertension are at increased risk of umbilical cord knotting resulting in stillbirths? The question of when to begin fetal testing in patients with hypertension is extremely important. A retrospective analysis could be quite informative. We believe that the data presented by Pircon et al. should be reanalyzed, taking into account the number of tests performed at any given gestational age, with a more complete data presentation. Then justifiable conclusions could be drawn as to when antepartum testing should be initiated in the patient with hypertension. Avick C. Mitra, MD, and William N.? Herbert, MD Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, University of North Carolina School of Medicine, CB# 7570, Chapel Hill, NC 27599-7570
Reply To the Editors: We thank Mitra and Herbert for their interest in our study of antepartum testing in patients with hypertension. The answers to their questions can be found in the text of our original publication; however, we are happy to review the results again. Our objectives were to evaluate when fetal compromise occurs in this patient population and to make recommendations as to when to initiate antepartum testing. Contrary to the suggestion of Mitra and Herbert, the data were not analyzed as a percentage of abnormal antepartum tests per number of patients. We were not interested in the age-old issue of sensitivity or specificity of a testing method in patients with hypertension; rather, we were interested in when compromise occurs. We defined compromise by several parameters, one of which was a positive contraction stress test. Test results were recorded prospectively and were retrospectively queried by first positive test, not by the number of positive tests the patients had. In evaluating the data in this fashion we found that hypertensive patients with concomitant diagnoses of diabetes demonstrated compromise at an earlier gestational age when compared with hypertensive patients without diabetes. Subsequent management was not the topic of study. When patient management is modified because of an antepartum test, there is evidence of fetal compromise. To reiterate, we were interested in when this occurred. We all agree that follow-up of an abnormal test can be managed numerous ways. Although it seems that Mitra and Herbert have expressed an interest in our management protocol, that protocol would not affect our
Letters
295
results or our conclusions in this study. Consequently we do not believe it is appropriate to review that topic here. We recommended that testing in the diabetic patient with hypertension start at viability. As listed in our original publication, this was based on two observations. First, there were two patients with late second-trimester stillbirths and concomitant diagnoses of a true knot in the umbilical cord. It is difficult to attribute the losses solely to the knots in the cords because this is a common finding at the deliveries of normal, viable term infants. We did not suggest that there is an increased risk of cord knotting in diabetic patients with hypertension. Second, a clinician would prefer to initiate testing in a patient before fetal compromise. The earliest gestational age that intervention was necessary in the diabetic patient with hypertension was 28 weeks. Therefore testing would need to start at 26 to 27 weeks or near the time of viability. Richard Pircon, MD St. joseph's Medical Center, 5000 W. Chambers St., Milwaukee, WI 53210
Roger Freeman, MD Memorial Women's Hospital, 2801 Atlantic Ave., Long Beach, CA 90801
Legal thoughts on malpractice claims: Causes and prevention
To the Editors: As a professor of consumer law married to a physician, I read with great interest Ward's article (Ward CJ. Analysis of 500 obstetric and gynecologic malpractice claims: causes and prevention 1991; 165:298-306). I believe the article would have been improved and Kirkley'S comments muted by a change in the classification method. Ward divides the cases into "defensible" and "indefensible," and Kirkley has a number of problems with this categorization. So do I, but for a different reason. There is simply no division in the law which corresponds to the defensible-indefensible dichotomy. The law divides cases into three categories, not two. In the first the plaintiff has no case and the defendant wins on summary judgment. In the second the defendant has no case and the plaintiff wins on summary judgment. In the third the issue is presented to the jury for determination of the disputed facts. Given Ward's description of his method it would appear that most, though not all, of his indefensible cases fall into the second category, where the plaintiff might win on summary judgment. The defensible cases would be in the other two categories. However, even if a case can be made for a physician'S actions, it does not mean that the physician will win. Ward's article may give physicians the unwarranted belief that if they lose a "defensible" case, the system has somehow failed them presumably because of perjured expert testimony, idiot jurors, and corrupt lawyers. This is simply not the law, just as it is not the law that a bad outcome means that there was malpractice. The legal system recognizes that there are often factual disputes on which reasonable people can differ and
