- Email: [email protected]
A prospective, randomized comparison of the Pipelle endometrial sampling device with the Novak curette
A prospective, randomized comparison of the Pipelle endometrial sampling device with the Novak curette Thomas G. Stovall, MD, Frank W. Ling, MD, and Patrick L. Morgan, MD Memphis, Tennessee This prospective, randomized study compares sample adequacy, pain associated with endometrial biopsy, and correlation of endometrial histologic sampling with hysterectomy histologic results in specimens obtained by the Pipelle (Unimar Inc., Wilton, Conn.) device (N = 149) and the Novak (Miltex, Inc., Lake Success, N.Y.) curette (N = 126). During a 2-year period, patients with abnormal uterine bleeding seen in our ambulatory care facility were enrolled in this trial, unless they did not wish to participate or had a positive pregnancy test result. Patients in both groups were similar with respect to mean gravidity, parity, and menopausal status, although patients in the Novak group were older (43.4 vs 39.1 years, p = 0.005). Patients undergoing Novak biopsy had a mean pain score of 4.36 with 22 of 126 (17%) reporting severe pain, whereas patients undergoing Pipelle biopsy had a mean pain score of 3.21 with only 10 of 149 (6.7%) reporting severe pain (p < 0.05). These pain scores were not affected by menstrual day, gravidity, parity, or menopausal status. Insufficient tissue was reported in 12.8% of patients in the Pipelle group compared with 9.5% in the Novak group (p > 0.05). Fifty patients underwent subsequent hysterectomy. In 48 of 50 (96%), the pathologic results at hysterectomy were in agreement with the histologic findings at endometrial sampling. This clinical trial suggests that Pipelle biopsy appears to be as effective as the Novak curette in obtaining an adequate specimen for histologic analysis and is associated with less pain. (AM J OSSTET GYNECOL 1991 ;165:1287-9.)
Key words: Endometrial biopsy, Novak curette. Pipelle endometrial sampling device Endometrial sampling is an important diagnostic tool in gynecologic practice and is used in the evaluation of abnormal uterine bleeding, for endometrial dating, and in the assessment of endometrial histologic results in women receiving hormonal therapy. Traditionally. curettage has been the gold standard sampling technique, although it is being replaced by simpler, less morbid, and less expensive office biopsy techniques. l The Novak (Miltex, Inc., Lake Success, N.Y.) curette and vacuum aspirator have been shown to be as accurate as curettage when endometrial histologic results are compared." The Pipelie (Unimar Inc., Wilton, Conn.) endometrial suction curette was developed as a method of endometrial sampling, causing less patient discomfort while obtaining adequate tissue for histologic diagnosis. To date, few comparative trials have been conducted comparing this sampling device with other, proved methods of endometrial sampling. 3 . 9 The purpose of this study is to compare sample adequacy and pain associated with sampling and to correlate endometrial sampling histologic findings. with
From the Division of Gynecology, Department of Obstet1'ics and G.vnecology, University of Tennessee, Memphis. Receit'edfor publication December 2 7, 1990; revised April 30, 1991; accepted May 15, 1991. Reprint requests: Thomas G. Stovall, MD, Department of Obstetrics and Gynecology, University of Tennessee, Memphis, 853 Jefferson Ave., Room E-102, Memphis, TN 38163.
6/1/31188
hysterectomy histologic results in specimens obtained by the Pipelle and Novak curettes.
Material and methods During a 2-year period ending February 1990, women with abnormal uterine bleeding seen at the Gailor Clinic, the ambulatory care facility of the University of Tennessee, Memphis, were enrolled in this prospective trial. Only women with a positive pregnancy test result or not wishing to participate were excluded. Patients were randomized according to computer-generated numbers to undergo endometrial sampling with either the Novak curette" (group 1) or the Pipelie (group 2) endometrial suction curette. The Pipelle device is made of a clear, flexible polypropylene sheath. It has a blunt, rounded distal tip and measures 23.5 cm in length. It has inner and outer diameters of 2.6 and 3.1 mm, respectively. By withdrawing on the inner piston, a negative pressure is created within the endometrial cavity. The Novak curette is a stainless steel curette with an outer diameter of 5 mm. Serrated edges surround its distal opening. A 10 ml plastic syringe was used to create a negative pressure. The position of the uterus was determined by pelvic examination, and the biopsy was performed after the cervix was exposed and cleaned with a povidone-iodine solution. No premedication was used for any patient. The Novak curette was inserted into the endometrial 1287
1288 Stovall, Ling, and Morgan
November 1991 Am J Obstet Gynecol
Table I. Comparison of demographic data and sampling between Novak and Pipelle groups Novak group (N
Age (yr) Median Range Gravidity Median Range Parity Median Range Biopsy indication Premenopausal abnormal uterine bleeding No.
= 126)
Pipelle group (N
= U9)
p Value
44 21-80
40 25-66
0.0005
5 0-17
4 0-25
NS
4 0-17
4 0-18
NS
%
105 83.3
134 89.9
NS
%
21 89.9
15 10.1
NS
Postmenopausal bleeding No.
NS, Not significant;
p ~ 0.05.
cavity and the specimen was obtained from the fundal portion. The Pipelle device was introduced into the endometrial cavity, and the piston was then withdrawn to the full length of the instrument, thereby creating a negative pressure. As the instrument was withdrawn, it was rotated, drawing tissue into the lumen. After the Pipelle device was removed, the tip was cut off and the sample expelled. All biopsy specimens were placed in 10% formaldehyde and sent for tissue processing and staining. The pathologist interpreting the endometrial samples was blinded to the instrument used to obtain the sample. The patient's age, gravidity, parity, and menstrual history were noted. Outcomes measured included specimen adequacy, i.e., the amount of tissue necessary for the pathologist to make a diagnosis. An adequate sample was defined as one or more pieces of endometrium large enough to determine the gland-to-stroma ratio and endometrial morphologic features. An inadequate sample was defined as consisting only of blood or cervical mucus with fragments of benign endocervix or a large amount of blood with only small fragments of endometrial glands and stroma. In those patients requiring hysterectomy, pathologic findings on endometrial biopsy were compared with endometrial histologic results at the time of hysterectomy. After the biopsy, the patient was asked to quantify the intensity of any pain experienced during the procedure. Pain was graded on a scale of 0 to 10, with 0 being no pain and 10 being severe pain with cramping. Statistical analysis was performed with a two-tailed Student t test and two-tailed Fisher's exact test, with a p value of <0.05 considered significant. Results
Two hundred seventy-five patients were enrolled in this study, 149 (54.2%) undergoing Pipelle device en-
dometrial sampling and 126 (45.8%) having Novak curette sampling. No uterine perforations or other complications were noted in either group. Table I compares the two groups with respect to patient age, gravidity, and parity. Patients in the Novak group were statistically older (p = 0.005) than those in the Pipelle group. Thirty-six (13.1 %) of the patients were postmenopausal. The inadequacy rate in this group of patients was 4 of 18 (22.2%) in the Pipelle group and 5 of 18 (27.8%) in the Novak group. Table I also details the indications for endometrial sampling in the two groups. There were no significant differences between groups. Patients undergoing Novak endometrial sampling had a mean pain score of 4.36, with 22 of 126 (17%) reporting a pain score of >7 (severe pain) and only 39 of 126 (31 %) reporting a score of <3 (minimal pain). Patients undergoing Pipelle endometrial sampling had a mean pain score of3.21, with 10 of 149 (7%) reporting a pain score of >7 (severe pain), whereas 60 of 149 (40%) reported a pain score of <3 (minimal pain). Patients were further subdivided by menstrual days, gravidity, parity, and menopausal status. Those reporting the least pain were in the Pipelle group during the second menstrual week, whereas those reponing the most pain were characterized by Novak biopsy after the third menstrual week. Although there was a significant difference between the two groups (p < 0.05) with regard to pain induced by biopsy, there was no difference within each group with regard to timing of the biopsy during the menstrual cycle. The histologic results of the biopsies are outlined in Table II. Although no patient had a failed endometrial biopsy, insufficient tissue for histologic diagnosis was obtained in 19 (12.8%) patients in the Pipelle group compared with 12 (9.5%) in the Novak group (p> 0.05). Patients in whom sufficient tissue for histologic analysis was not obtained subsequently either
Endometrial sampling
Volume 165 Number 5, Part I
1289
Table II. Comparison of endometrial histologic results between Novak and Pipelle groups Endometrial histologic finding Sampling method Pipelle (N No.
=
ovak (N No.
=
% %
149) 126)
Endometritis
Hyperplasia
Proliferative or secretory
Insufficient tissue
23 15.4
II 7.4
96 64.4
19 12.8
23 18.3
15 11.9
76 60.3
12 9.5
had repeat endometrial sampling or underwent curettage. No patient in this group was found to have either endometrial hyperplasia or carcinoma. Fifty patients underwent hysterectomy subsequent to endometrial sampling, 26 (17%) in the Pipelle group and 24 (19%) in the Novak group. In 48 of 50 (96%) patients, the endometrial histologic results at tlH: time of hysterectomy were in agreement with the endometrial sampling histologic results. The hysterectomy specimens of one patient in each group differed significantly from the biopsy findings. In the Novak group a mixed miillerian tumor was found, whereas endometrial hyperplasia without atypicality was found in the Pipelie group. In both patients endometrial sampling had revealed proliferative endometrium. Comment
This randomized clinical trial suggests that the Pipelle device is as effective as the Novak curette in obtaining endometrial tissue for histologic evaluation. Both the adequacy and quality of the samples obtained by the two methods appeared similar. The study had a 0.129 power of detecting a difference in inadequacy rates of 12.8% and 9.5% between the Pipelle and Novak groups, respectively. To detect a similar difference at a power of 0.80, a sample size of 1507 patients would be required in each group. Thus no definite conclusion can be made in this regard. Unlike other trials,1O this patient population was made up of 13.1 % of women who were postmenopausal. The inadequacy rates in postmenopausal patients were 22.2% and 27.8% in the Pipelie and Novak groups, respectively. In this group of older patients, adequate tissue for histologic analysis was obtained in 27 of 36 patients, thereby eliminating the risk and cost of curettage for these patients. Because the Pipelie has a smaller diameter, there is less difficulty in passing the instrument through the cervix and therefore less pain. The smaller size also allows potential entry into uteri with relative stenotic cervices as seen in some menopausal patients. As a result, the Pipelle device may be a preferable instrument in selected postmenopausal patients. The Pipelle device was associated with less discomfort than was the Novak curette. This was most likely related
to its flexibility and smaller caliber. The Pipelie device can be made more rigid by placing it in the freezer, but in this study the Pipelle devices were not pretreated. Although the women in the Novak group were older, confounding is unlikely to have influenced the findings. When the results were reevaluated to eliminate age bias, the results were not significantly altered. In addition, chance also is an unlikely explanation for this difference. This study had a 0.898 power of detecting a difference in pain score as large as 3.21 versus 4.36. Although previous studies revealed the Pipelle to be well tolerated:·6 to our knowledge this study is the first to quantify patient discomfort. This decreased pain may be a result of the Pipelie's smaller diameter and flexible tip. There also may be an effect of the age differences found between the two groups. Two patients had an inaccurate diagnosis made at the time of endometrial biopsy. Because not all patients underwent hysterectomy, it is certainly possible that other lesions were not diagnosed. However, it is important to remember that even curettage does not sample the entire endometrium and will, on occasion, not identify a clinically important lesion. The single-use disposable Pipelle device costs $6 per unit. The initial cost of the nondisposable Novak curette is $31. Sterilization and packaging and the purchase of the required syringe are additional costs that are incurred with each use for the lifetime of the Novak instrument. However, because the Novak curette's expected useful lifetime is unlimited, whereas a new Pipelle device must be used for each biopsy, the cost factor favors the Novak curette as the number of biopsies performed increases. In fact, depending on sterilization and packaging costs, the Pipelle device may result in increased costs over the Novak curette. The cost of processing the tissue sample and histologic interpretation is the same for both methods. REFERENCES 1. Grimes DA. Diagnostic dilation and curettage: a reappraisal. AMJ OBSTET GYNECOL 1982;142:1-6. 2. Stovall TG, Solomon SK, Ling FW. Endometrial sampling prior to hysterectomy. Obstet Gynecol 1989;73:405-9. 3. Cornier E. The Pipelle: a disposable device for endometrial biopsy. AM J OBSTET GYNECOL 1984; 148: 109-10.
November 1991 Am J Obstet Gynecol
Stovall, Ling, and Morgan
4. Kaunitz AM, Masciello A, Ostrowski M. Rovira EZ. Comparison of endometrial Pipelle and Vabra aspirator. J Reprod Med 1988;33:427-31. 5. Henig I, Chan P, Tredway DR, Maw GM, Gullett AJ, Cheatwood M. Evaluation of Pipelie curette for endometrial biopsy. J Reprod Hed 1989;34:786-9. 6. Hill GA. Herbert CM Ill, Parker RA, Wentz AC. Comparison of late luteal phase endometrial biopsies using the Novak curette or Pipelle endometrial suction curette. Obstet Gynecol 1989;73:443-5. 7. CheckJH, Chase'/S, Nowroozi K. Wu CH. Chern R. Clin-
ical evaluation of the Pipelle endometrial suction curette for timed endometrial biopsies. ./ Reprod Med 1989;34:218-20.
8. Bibbo M, Kluskens L, Azizi F, et al. Accuracy of three sampling techniques for the diagnosis of endometrial cancer and hyperplasia. ./ Reprod Med 1982;27:622-6. 9. Koonings PP, Moyer DDDL, Grimes DA. A randomized clinical trial comparing Pipelle and Tis-V-Trap for endometrial biopsy. Obstet Gynecol 1990;75:293-5. 10. Novak E. A suction-curet apparatus for endometrial biopsy. .lAMA 1935;104:1497-8.
Spontaneous preterm birth: A case-control study Ine de Haas, MD, Bernard L. Harlow, PhD, Daniel W. Cramer, MD, SeD, and Fredric D. Frigoletto, Jr., MD
Boston, Massachusetts Information on demographic characteristics, life-style, and reproductive, prenatal, and medical history was obtained from the prenatal records of 140 women with a preterm delivery (cases) and 280 women with a term delivery (controls) matched by age. Compared with controls, a greater proportion of cases were nonwhite, never married, and educated <12 years. When logistic regression was used to control for confounding variables, a number of risk factors were found to be associated with an increased risk for preterm delivery: a history of a prior preterm delivery (relative risk, 3.5; confidence interval, 1.6 to 7.8), smoking during the pregnancy (p value test of trends, 0.005), and a prepregnancy weight <61.5 kg (relative risk, 2.0; 95% confidence interval, 1.2 to 3.2). Seven (5%) cases and none of the controls indicated a history of maternal diethylstilbestrol exposure (p < 0.001). A history of prior induced abortion was associated with a modest increase in risk for a spontaneous preterm delivery (relative risk, 1.6; confidence interval, 0.9 to 2.7), and this risk increased with increasing numbers of induced abortions (p value for x2 test of trend, 0.02). (AM J OSSTET GVNECOL 1991 ;165:1290-6.)
Key words: PreteI'm delivery, case-control study, obstetric history, cigarette smoking, maternal weight In the United States, preteI'm delivery. defined as the birth of a fetus before 37 completed weeks of gestation, occurs in about 9% of all single live births and accounts for a large proportion of neonatal morbidity and mortality.' Approximately two thirds of all preteI'm births occur spontaneously after premature labor or premature rupture of the membranes. 2 The remaining preteI'm births are due to adverse maternal or fetal diagnoses that warrant an early delivery, such as hypertension, antepartum hemorrhage, diabetes, or fetal growth
From the Obstetrics and Gynecology Epidemiology Center, Brigham and Women's Hospital, Harvard Medical School. Receivedfor publication December 11, 1990; revised April 26. 1991; accepted May 6,1991. Reprint requests: Bernard L. Ha1'low, PhD, Ob/Gyn Epidemiology Center, BTigham and Women's Hospital, 221 Longwood Ave.. Boston, MA 02115. 6/1 /30704 12~O
retardation.' This report seeks to examine risk factors for spontaneous preteI'm delivery that could be identified from the prenatal history. and our results are based on a case-control study in which medical records were used.
Methods Cases were defined as women who were delivered of a live singleton infant between 20 and 37 weeks' gestation and in whom the delivery had been preceded by spontaneous labor or rupture of the membranes without induction for maternal or fetal indications. Cases were selected from all deliveries between Oct. 1, 1988, and March 31, 1989, at Brigham and Women's Hospital. We identified 263 deliveries that started spontaneously before 37 weeks (ICD-9-CM codes 644.2. 658.1. and 658.2). Gestational age at delivery was defined by last menstrual period (LMP) confirmed by clin-
Key words: Endometrial biopsy, Novak curette. Pipelle endometrial sampling device Endometrial sampling is an important diagnostic tool in gynecologic practice and is used in the evaluation of abnormal uterine bleeding, for endometrial dating, and in the assessment of endometrial histologic results in women receiving hormonal therapy. Traditionally. curettage has been the gold standard sampling technique, although it is being replaced by simpler, less morbid, and less expensive office biopsy techniques. l The Novak (Miltex, Inc., Lake Success, N.Y.) curette and vacuum aspirator have been shown to be as accurate as curettage when endometrial histologic results are compared." The Pipelie (Unimar Inc., Wilton, Conn.) endometrial suction curette was developed as a method of endometrial sampling, causing less patient discomfort while obtaining adequate tissue for histologic diagnosis. To date, few comparative trials have been conducted comparing this sampling device with other, proved methods of endometrial sampling. 3 . 9 The purpose of this study is to compare sample adequacy and pain associated with sampling and to correlate endometrial sampling histologic findings. with
From the Division of Gynecology, Department of Obstet1'ics and G.vnecology, University of Tennessee, Memphis. Receit'edfor publication December 2 7, 1990; revised April 30, 1991; accepted May 15, 1991. Reprint requests: Thomas G. Stovall, MD, Department of Obstetrics and Gynecology, University of Tennessee, Memphis, 853 Jefferson Ave., Room E-102, Memphis, TN 38163.
6/1/31188
hysterectomy histologic results in specimens obtained by the Pipelle and Novak curettes.
Material and methods During a 2-year period ending February 1990, women with abnormal uterine bleeding seen at the Gailor Clinic, the ambulatory care facility of the University of Tennessee, Memphis, were enrolled in this prospective trial. Only women with a positive pregnancy test result or not wishing to participate were excluded. Patients were randomized according to computer-generated numbers to undergo endometrial sampling with either the Novak curette" (group 1) or the Pipelie (group 2) endometrial suction curette. The Pipelle device is made of a clear, flexible polypropylene sheath. It has a blunt, rounded distal tip and measures 23.5 cm in length. It has inner and outer diameters of 2.6 and 3.1 mm, respectively. By withdrawing on the inner piston, a negative pressure is created within the endometrial cavity. The Novak curette is a stainless steel curette with an outer diameter of 5 mm. Serrated edges surround its distal opening. A 10 ml plastic syringe was used to create a negative pressure. The position of the uterus was determined by pelvic examination, and the biopsy was performed after the cervix was exposed and cleaned with a povidone-iodine solution. No premedication was used for any patient. The Novak curette was inserted into the endometrial 1287
1288 Stovall, Ling, and Morgan
November 1991 Am J Obstet Gynecol
Table I. Comparison of demographic data and sampling between Novak and Pipelle groups Novak group (N
Age (yr) Median Range Gravidity Median Range Parity Median Range Biopsy indication Premenopausal abnormal uterine bleeding No.
= 126)
Pipelle group (N
= U9)
p Value
44 21-80
40 25-66
0.0005
5 0-17
4 0-25
NS
4 0-17
4 0-18
NS
%
105 83.3
134 89.9
NS
%
21 89.9
15 10.1
NS
Postmenopausal bleeding No.
NS, Not significant;
p ~ 0.05.
cavity and the specimen was obtained from the fundal portion. The Pipelle device was introduced into the endometrial cavity, and the piston was then withdrawn to the full length of the instrument, thereby creating a negative pressure. As the instrument was withdrawn, it was rotated, drawing tissue into the lumen. After the Pipelle device was removed, the tip was cut off and the sample expelled. All biopsy specimens were placed in 10% formaldehyde and sent for tissue processing and staining. The pathologist interpreting the endometrial samples was blinded to the instrument used to obtain the sample. The patient's age, gravidity, parity, and menstrual history were noted. Outcomes measured included specimen adequacy, i.e., the amount of tissue necessary for the pathologist to make a diagnosis. An adequate sample was defined as one or more pieces of endometrium large enough to determine the gland-to-stroma ratio and endometrial morphologic features. An inadequate sample was defined as consisting only of blood or cervical mucus with fragments of benign endocervix or a large amount of blood with only small fragments of endometrial glands and stroma. In those patients requiring hysterectomy, pathologic findings on endometrial biopsy were compared with endometrial histologic results at the time of hysterectomy. After the biopsy, the patient was asked to quantify the intensity of any pain experienced during the procedure. Pain was graded on a scale of 0 to 10, with 0 being no pain and 10 being severe pain with cramping. Statistical analysis was performed with a two-tailed Student t test and two-tailed Fisher's exact test, with a p value of <0.05 considered significant. Results
Two hundred seventy-five patients were enrolled in this study, 149 (54.2%) undergoing Pipelle device en-
dometrial sampling and 126 (45.8%) having Novak curette sampling. No uterine perforations or other complications were noted in either group. Table I compares the two groups with respect to patient age, gravidity, and parity. Patients in the Novak group were statistically older (p = 0.005) than those in the Pipelle group. Thirty-six (13.1 %) of the patients were postmenopausal. The inadequacy rate in this group of patients was 4 of 18 (22.2%) in the Pipelle group and 5 of 18 (27.8%) in the Novak group. Table I also details the indications for endometrial sampling in the two groups. There were no significant differences between groups. Patients undergoing Novak endometrial sampling had a mean pain score of 4.36, with 22 of 126 (17%) reporting a pain score of >7 (severe pain) and only 39 of 126 (31 %) reporting a score of <3 (minimal pain). Patients undergoing Pipelle endometrial sampling had a mean pain score of3.21, with 10 of 149 (7%) reporting a pain score of >7 (severe pain), whereas 60 of 149 (40%) reported a pain score of <3 (minimal pain). Patients were further subdivided by menstrual days, gravidity, parity, and menopausal status. Those reporting the least pain were in the Pipelle group during the second menstrual week, whereas those reponing the most pain were characterized by Novak biopsy after the third menstrual week. Although there was a significant difference between the two groups (p < 0.05) with regard to pain induced by biopsy, there was no difference within each group with regard to timing of the biopsy during the menstrual cycle. The histologic results of the biopsies are outlined in Table II. Although no patient had a failed endometrial biopsy, insufficient tissue for histologic diagnosis was obtained in 19 (12.8%) patients in the Pipelle group compared with 12 (9.5%) in the Novak group (p> 0.05). Patients in whom sufficient tissue for histologic analysis was not obtained subsequently either
Endometrial sampling
Volume 165 Number 5, Part I
1289
Table II. Comparison of endometrial histologic results between Novak and Pipelle groups Endometrial histologic finding Sampling method Pipelle (N No.
=
ovak (N No.
=
% %
149) 126)
Endometritis
Hyperplasia
Proliferative or secretory
Insufficient tissue
23 15.4
II 7.4
96 64.4
19 12.8
23 18.3
15 11.9
76 60.3
12 9.5
had repeat endometrial sampling or underwent curettage. No patient in this group was found to have either endometrial hyperplasia or carcinoma. Fifty patients underwent hysterectomy subsequent to endometrial sampling, 26 (17%) in the Pipelle group and 24 (19%) in the Novak group. In 48 of 50 (96%) patients, the endometrial histologic results at tlH: time of hysterectomy were in agreement with the endometrial sampling histologic results. The hysterectomy specimens of one patient in each group differed significantly from the biopsy findings. In the Novak group a mixed miillerian tumor was found, whereas endometrial hyperplasia without atypicality was found in the Pipelie group. In both patients endometrial sampling had revealed proliferative endometrium. Comment
This randomized clinical trial suggests that the Pipelle device is as effective as the Novak curette in obtaining endometrial tissue for histologic evaluation. Both the adequacy and quality of the samples obtained by the two methods appeared similar. The study had a 0.129 power of detecting a difference in inadequacy rates of 12.8% and 9.5% between the Pipelle and Novak groups, respectively. To detect a similar difference at a power of 0.80, a sample size of 1507 patients would be required in each group. Thus no definite conclusion can be made in this regard. Unlike other trials,1O this patient population was made up of 13.1 % of women who were postmenopausal. The inadequacy rates in postmenopausal patients were 22.2% and 27.8% in the Pipelie and Novak groups, respectively. In this group of older patients, adequate tissue for histologic analysis was obtained in 27 of 36 patients, thereby eliminating the risk and cost of curettage for these patients. Because the Pipelie has a smaller diameter, there is less difficulty in passing the instrument through the cervix and therefore less pain. The smaller size also allows potential entry into uteri with relative stenotic cervices as seen in some menopausal patients. As a result, the Pipelle device may be a preferable instrument in selected postmenopausal patients. The Pipelle device was associated with less discomfort than was the Novak curette. This was most likely related
to its flexibility and smaller caliber. The Pipelie device can be made more rigid by placing it in the freezer, but in this study the Pipelle devices were not pretreated. Although the women in the Novak group were older, confounding is unlikely to have influenced the findings. When the results were reevaluated to eliminate age bias, the results were not significantly altered. In addition, chance also is an unlikely explanation for this difference. This study had a 0.898 power of detecting a difference in pain score as large as 3.21 versus 4.36. Although previous studies revealed the Pipelle to be well tolerated:·6 to our knowledge this study is the first to quantify patient discomfort. This decreased pain may be a result of the Pipelie's smaller diameter and flexible tip. There also may be an effect of the age differences found between the two groups. Two patients had an inaccurate diagnosis made at the time of endometrial biopsy. Because not all patients underwent hysterectomy, it is certainly possible that other lesions were not diagnosed. However, it is important to remember that even curettage does not sample the entire endometrium and will, on occasion, not identify a clinically important lesion. The single-use disposable Pipelle device costs $6 per unit. The initial cost of the nondisposable Novak curette is $31. Sterilization and packaging and the purchase of the required syringe are additional costs that are incurred with each use for the lifetime of the Novak instrument. However, because the Novak curette's expected useful lifetime is unlimited, whereas a new Pipelle device must be used for each biopsy, the cost factor favors the Novak curette as the number of biopsies performed increases. In fact, depending on sterilization and packaging costs, the Pipelle device may result in increased costs over the Novak curette. The cost of processing the tissue sample and histologic interpretation is the same for both methods. REFERENCES 1. Grimes DA. Diagnostic dilation and curettage: a reappraisal. AMJ OBSTET GYNECOL 1982;142:1-6. 2. Stovall TG, Solomon SK, Ling FW. Endometrial sampling prior to hysterectomy. Obstet Gynecol 1989;73:405-9. 3. Cornier E. The Pipelle: a disposable device for endometrial biopsy. AM J OBSTET GYNECOL 1984; 148: 109-10.
November 1991 Am J Obstet Gynecol
Stovall, Ling, and Morgan
4. Kaunitz AM, Masciello A, Ostrowski M. Rovira EZ. Comparison of endometrial Pipelle and Vabra aspirator. J Reprod Med 1988;33:427-31. 5. Henig I, Chan P, Tredway DR, Maw GM, Gullett AJ, Cheatwood M. Evaluation of Pipelie curette for endometrial biopsy. J Reprod Hed 1989;34:786-9. 6. Hill GA. Herbert CM Ill, Parker RA, Wentz AC. Comparison of late luteal phase endometrial biopsies using the Novak curette or Pipelle endometrial suction curette. Obstet Gynecol 1989;73:443-5. 7. CheckJH, Chase'/S, Nowroozi K. Wu CH. Chern R. Clin-
ical evaluation of the Pipelle endometrial suction curette for timed endometrial biopsies. ./ Reprod Med 1989;34:218-20.
8. Bibbo M, Kluskens L, Azizi F, et al. Accuracy of three sampling techniques for the diagnosis of endometrial cancer and hyperplasia. ./ Reprod Med 1982;27:622-6. 9. Koonings PP, Moyer DDDL, Grimes DA. A randomized clinical trial comparing Pipelle and Tis-V-Trap for endometrial biopsy. Obstet Gynecol 1990;75:293-5. 10. Novak E. A suction-curet apparatus for endometrial biopsy. .lAMA 1935;104:1497-8.
Spontaneous preterm birth: A case-control study Ine de Haas, MD, Bernard L. Harlow, PhD, Daniel W. Cramer, MD, SeD, and Fredric D. Frigoletto, Jr., MD
Boston, Massachusetts Information on demographic characteristics, life-style, and reproductive, prenatal, and medical history was obtained from the prenatal records of 140 women with a preterm delivery (cases) and 280 women with a term delivery (controls) matched by age. Compared with controls, a greater proportion of cases were nonwhite, never married, and educated <12 years. When logistic regression was used to control for confounding variables, a number of risk factors were found to be associated with an increased risk for preterm delivery: a history of a prior preterm delivery (relative risk, 3.5; confidence interval, 1.6 to 7.8), smoking during the pregnancy (p value test of trends, 0.005), and a prepregnancy weight <61.5 kg (relative risk, 2.0; 95% confidence interval, 1.2 to 3.2). Seven (5%) cases and none of the controls indicated a history of maternal diethylstilbestrol exposure (p < 0.001). A history of prior induced abortion was associated with a modest increase in risk for a spontaneous preterm delivery (relative risk, 1.6; confidence interval, 0.9 to 2.7), and this risk increased with increasing numbers of induced abortions (p value for x2 test of trend, 0.02). (AM J OSSTET GVNECOL 1991 ;165:1290-6.)
Key words: PreteI'm delivery, case-control study, obstetric history, cigarette smoking, maternal weight In the United States, preteI'm delivery. defined as the birth of a fetus before 37 completed weeks of gestation, occurs in about 9% of all single live births and accounts for a large proportion of neonatal morbidity and mortality.' Approximately two thirds of all preteI'm births occur spontaneously after premature labor or premature rupture of the membranes. 2 The remaining preteI'm births are due to adverse maternal or fetal diagnoses that warrant an early delivery, such as hypertension, antepartum hemorrhage, diabetes, or fetal growth
From the Obstetrics and Gynecology Epidemiology Center, Brigham and Women's Hospital, Harvard Medical School. Receivedfor publication December 11, 1990; revised April 26. 1991; accepted May 6,1991. Reprint requests: Bernard L. Ha1'low, PhD, Ob/Gyn Epidemiology Center, BTigham and Women's Hospital, 221 Longwood Ave.. Boston, MA 02115. 6/1 /30704 12~O
retardation.' This report seeks to examine risk factors for spontaneous preteI'm delivery that could be identified from the prenatal history. and our results are based on a case-control study in which medical records were used.
Methods Cases were defined as women who were delivered of a live singleton infant between 20 and 37 weeks' gestation and in whom the delivery had been preceded by spontaneous labor or rupture of the membranes without induction for maternal or fetal indications. Cases were selected from all deliveries between Oct. 1, 1988, and March 31, 1989, at Brigham and Women's Hospital. We identified 263 deliveries that started spontaneously before 37 weeks (ICD-9-CM codes 644.2. 658.1. and 658.2). Gestational age at delivery was defined by last menstrual period (LMP) confirmed by clin-