- Email: [email protected]
The integrated luteal progesterone: an assessment of luteal function
Vol. 48, No.6, December 1987 Printed in U.S.A.
FERTILITY AND STERILITY Copyright c 1987 The American Fertility Society
The integrated luteal progesterone: an assessment of luteal function
Chung H. Wu, M.D.* Shahab S. Minassian, M.D.t Jefferson Medical College of Thomas Jefferson University, and The Medical College of Pennsylvania, Philadelphia, Pennsylvania
An integrated luteal progesterone (ILP L ) was calculated on the basis of a luteal progesterone (P) level with the assumption that the daily plasma P level in the luteal phase closely approximates a sine curve. The midluteal P-amplitude (K) was also obtained mathematically. Daily luteal P levels from five normal ovulatory cycles were assessed for the biologic variation of ILP L and K, then compiled to construct a normogram of the ILP during the luteal phase. The coefficient of variation of K and total ILP L in each cycle ranged from 9.7% to 24.3% and 3.5% to 13.2%, respectively. Fifty-two infertility patients were evaluated for their luteal function by the luteal P and estradiol (E 2) level, K, ILPL , endometrial biopsy (EBX)-lag-day, as well as the lengths of follicular phase, luteal phase (L#), and cycle. Thirty-nine patients had EBX-lag day:$; 2 days and were designated as infertile-normal (INF-NL) luteal phase, while the remaining 13 patients who had EBXlag day> 2 days were considered as luteal phase defect (LPD). Significant (P < 0.05) differences were observed between INF-NL and LPD in: luteal length (13.2 ± 0.31 versus 11.0 ± 0.58 days, respectively), and total ILP L (170 ± 8.3 versus 113 ± 8.5 ng/ml-day, respectively). No differences were seen in luteal P, E2 and K levels, nor in follicular and cycle length. Significant (P < 0.05) correlations were observed between total ILP L and luteal P, E 2, L#, and K; while a negative correlation was noted between follicular and luteal length. Briefly, the total ILPL is a convenient and reliable parameter to assess luteal function in addition to or in place of EBX; and it can be used in clinical research or in monitoring therapy for ovulatory dysfunction. Fertil Steri148:937, 1987
The adequacy of luteal function traditionally has been evaluated most often by the late luteal phase endometrial biopsy (EBX) and, less frequently, by progesterone (P) levels in the serum or plasma and/or luteal length by basal body temperature (BBT). While the EBX is the most widely accepted criterion for luteal phase defect (LPD), it is not without limitations. Criteria for normal versus abnormal readings differ/ large interpersonal varia-
Received March 3, 1987; revised and accepted June 17, 1987. H. Wu, M.D., Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, Thomas Jefferson University, 1025 Walnut Street, Philadelphia, Pennsylvania 19107. t Department of Obstetrics and Gynecology, The Medical College of Pennsylvania.
* Reprint requests: Chung
Vol. 48, No.6, December 1987
tions have been reported, and its accuracy has been questioned. 2 ,3 The expense and discomfort of multiple EBX for diagnosis and monitoring of therapy for LPD is yet another problem. P determinations most accurately reflect luteal P secretion only if clinically impractical daily samples are obtained, and the use of one or three midluteallevels is controversia1. 4 ,5 BBT has been shown to be of some use in the diagnosis and treatment of LPD, 6 but cannot be used alone in this regard. There is a great need therefore for a reliable parameter to assess luteal function in the clinical and clinical research setting. The typical pattern of daily luteal phase P level in the nonpregnant state shows a distinct similarity to a sine curve. The area under the curve of individual cycle was integrated based on the luteal P level Wu and Minassian Integrated luteal progesterone
937
and the luteal length and was designated as the "integrated luteal progesterone" (ILPd. The ILPL then was applied to evaluate the EBX-proven normal and luteal phase defective cycles to assess its usefulness in the clinical evaluation of luteal function.
at exact midpoint of the sine curve. Assigning the luteal length as "L" and the day of blood sampling as "D", then: P = K sin (1rD/L) ILPD =
i i D
P=
D
K sin (1rD/L)
MATERIALS AND METHODS
=
when D
Normal Ovulatory Cycles
Five normally menstruating women with 27- to 30-day intervals were recruited after the proven normalcy of the basic endocrine study and physical examination. The daily blood samples were obtained and analyzed for luteinizing hormone (LH), P, and estradiol (E 2). The data were synchronized at LH peak as day o. The BBT also was obtained as an additional reference to assess the luteal phase length. The end of cycles was designated as the beginning of the next menstruation. Mathematic Formulae for Integrated Luteal Progesterone and Progesterone-Amplitude Level
The P-amplitude level (K) is the largest P level calculated on the sine curve (Fig. 1), which occurs 18
16
r:zzl ILPD o ILPt.
t DAY
Figure 1 The mean (±SD) luteal progesterone (e e) levels from five normal cycles superimposed on a sine curve (0 0). K, P-amplitude; ILPD , integrated luteal progesterone up to the day uD"; ILP L , total integrated luteal progesterone. Formulae: for calculation of K, ILP D and ILPL from P level and the luteal length based on the formula of sine curve.
938
i.e., K = P/sin (1rD/L)
Wu and Minassian Integrated luteal progesterone
=
[1 - cos (1I"D/L)]KLhr
L, ILP L
=
2KL/1I"
where ILPD is the area under the sine curve up to the day "D" and ILPL is the total area under the sine curve that represents the total luteal P secretion. Using these formulae, the K, ILP D , and ILP L were calculated from the daily P data in the luteal phase of five normal cycles. The data from the first and last two days of the luteal phase were excluded from evaluation because they tended to produce higher error. The variation of K and ILPL so-obtained were assessed by the coefficient of variation (CV) (i.e., CV = standard deviation/mean). Study Patients
The luteal phases of 52 infertile women were evaluated during an infertility workup by BBT, EBX, midluteal P, and E2 levels drawn in the morning. Luteal length was determined by BBT, starting on the day when temperature rose above the midpoint between the mean levels of the follicular and luteal phase temperature. EBX were obtained during the late luteal phases of these cycles and dated by the criteria of Noyes et aU Single luteal blood samples were obtained during the middle two thirds of the luteal phase and were measured for P and E2 by standard radioimmunoassay techniques. A diagnosis of a LPD was made when a greater than 2-day lag was noted. Those cycles in which a less than or equal to 2-day lag was seen were designated as normal infertile (INF-NL) cycles. It must be noted that the diagnosis of a single luteal phase defective cycle does not mean that the diagnosis was made of LPD itself as a cause for the infertility. Statistical analysis was performed using Student's t-test. Correlation analysis was obtained by calculating correlation coefficients (r) between paired parameters in all study patients. RESULTS
The mean daily plasma P levels from the luteal phases of five normal cycles are shown in Figure 1. Fertility and Sterility
These women had regular menses, no clinical signs of androgen excess, and normal endocrine profiles. As shown in Figure 1, the daily luteal P level approximates a sine curve and is superimposed on an actual sine curve. The calculated mean values of the K and ILPL from these cycles ranged as follows: K, 14.4 to 20.5 with total mean 16.4 ng/ml; and ILPL , 128 to 185 with total mean 147 ng/ml/day. The coefficients of variation of K and ILP L for individual cycles ranged from 9.7% to 24.3% and 3.5% to 13.2%, respectively, and increased to 26.2% and 25.8%, respectively, when all cycles were combined. The mean daily cumulative ILP n for these cycles is shown in Figure 2. The shaded area includes the mean ILP n ± 1 standard deviation (SD). The portion of the graph below the shaded area represents inadequate P secretion, while the area above represents adequate secretion. Of the infertile patient cycles, 39 cycles were considered normal, and 13 cycles were diagnosed as LPD by EBX. The calculated ILPL and K, along with the midluteal P and E2 levels, and the length of follicular phase, luteal phase, and cycle are shown in Table 1. The mean ILPL of LPD cycles was significantly (P < 0.01) lower than in INF-NL cycles. There was no difference, however, in mean P level between groups. When cycles in which P levels were drawn on the third or fourth day prior to menses were analyzed, the ILP L of LPD cycles were again significantly (P < 0.01) lower than INF-NL. The luteal phase length of LPD cycles
110
110
Table 1 The Results of Ovarian Function Assessment in the Infertile-Normal (INF-NL) and the Luteal Phase Defect (LPD) Cycles Parameter
Unit
INF-NL (n = 39)
LPD (n = 13)
Midluteal P Midluteal E2
ng/ml pg/ml
14.6 ± Loa 157 ± 19
EBX lag-day
day
0.35 ± 0.11
3.1 ± 0.39 b
Follicular length Luteal length Cycle length
day day day
14.6 ± 0.44 13.2 ± 0.31 27.5 ± 0.36
15.5 ± 0.77 11.0 ± 0.58 b 26.5 ± 0.71
P-amplitude (K) ILPL
ng/ml ng/ml/day
19.8 ± 1.3 170 ± 8.3
18.0 ± 1.6 113 ± 8.5 b
10.5 ± 2.0 128 ± 11
Samples drawn on the 3rd or 4th day before menses
Midluteal P ILPL a b
ng/ml ng/ml/day
(n = 23)
(n = 8)
13.6 ± 1.2 171 ± 11.5
9.2 ± 2.8 104 ± 7.6 b
Mean ± SE. P < 0.01 when compared with the respective INF-NL value.
were significantly (P < 0.01) shorter than INF-NL cycles (11.0 ± 0.58 day versus 13.2 ± 0.31, respectively), but no difference in follicular phase length or cycle length was seen. When an ILPL cut-off of 130 ng/ml-day was used to evaluate the study patients, 11 of the 22 (50%) cycles below this level were false-positive as they were normal by EBX. Conversely, 2 of 30 (7%) cycles above this level were false-negative as they were LPD by EBX (sensitivity of93%). An analysis of levels taken on the third and fourth days prior to menses revealed essentially the same results. Correlation analysis was performed on the parameters assessed for ovarian function in the study patients. ILPL was significantly (P < 0.05) correlated with midluteal P (r = 0.662), E2 (r = 0.399), K (r = 0.875), and luteal length (r = 0.495); while K also was correlated with midluteal P (r = 0.681) and E2 (r = 0.401). Additional correlations were noted between cycle length and midluteal E2 (r = 0.422), follicular length (r = 0.736), and luteal length (r = 0.304); and midluteal E2 versus midluteal P (r = 0.439). A significant negative correlation was only observed between follicular length and luteal length (r = -0.532).
20
o
DISCUSSION
o
I.
~
Q
M 00
WTtAI. PHASE
Figure 2 The mean (±SD) cumulative integrated luteal progesterone (ILPD ) in the luteal phase of five normal cycles. Vol. 48, No.6, December 1987
Serum P levels have long been used as an indicator of normal and abnormal luteal function in the infertile woman; however, there has been much Wu and Minassian Integrated luteal progesterone
939
controversy over their value. Daily P levels taken throughout the entire luteal phase would be an ideal method but, obviously, is clinically impractical. Abraham et al5 reported on the use of three midluteal levels with good results, but this is an expensive and inconvenient option. Radwanska et al4 and Hammond and Talbert6 have advocated the use of a single midluteal P level, reporting good results. The advantages of less inconvenience, less cost, and clinical practicality are certainly clear, however, because of reported variations in the single level, the method has gained little general acceptance and the EBX remains the standard for LPD diagnosis. Although there is no denying the usefulness of the EBX as a luteal phase parameter, questions concerning its accuracy, high variation, the patient discomfort, and expense of multiple biopsies for the diagnosis and stepwise evaluation of therapy limits the use of this method in infertility therapy and clinical research. The integration of P levels during the luteal phase has been reported previously,S but no studies have yet proven the adequacy or inadequacy of these cycles by EBX or reported on the exact method of integration. Moreover, it relied on cumbersome daily P determinations. The secretion of P during the luteal phase approximates a sine curve, as evidenced by the five normal cycles in this study. By integrating and representing the entire curve of corpus luteum P secretion with one sample, the ILP L combines the reliability of daily measurements with the convenience of the single level. Also, by virtue of the great difference in the mean ILP L between biopsy-proven LPD and normal cycles, the ILPL appears to reflect what would be found on biopsies and may be used as an adjunct or even a replacement for the EBX when multiple biopsies are indicated. The ILPL also proves to be more reliable than single P levels because no difference in the mean P levels between groups was found in this study. Regardless of the convenience of the single P determination (for ILPd versus multiple levels, a separate vi~it to the laboratory or office would theoretically be required, and would mean two luteal phase visits if an EBX were performed in that cycle. An ideal clinical situation, therefore, would consist of a simultaneous EBX and P taken during
940
Wu and Minassian Integrated luteal progesterone
one visit. Because the EBX is accurate when performed during the late luteal phase, the ILPL results from this segment were examined. It was found that LPD cycles had a much lower mean ILPL than INF-NL cycles, suggesting that only one well-timed visit is necessary. To use the ILPL in an individual patient in the clinical setting, it would be helpful to establish a cutoff for normal levels. When the 130 ng/ml-day level was used for this purpose, 93% of normal ILPL agreed with EBX results. Conversely, nearly 50% of ILPL below 130 ng/ml-day were from INF-NL cycles. Should the clinician use this cutoff, therefore, it appears that, for individual cycles, a normal ILPL may be more meaningful than an abnormal one, and that a low ILPL level during infertility therapy may necessitate a confirmatory EBX. For research purposes, the mean ILPL does show quite a significant difference between groups, and can be used with assurance in this regard. In summary, the ILP L is a convenient, reliable test for luteal phase adequacy. By reflecting the entire secretion of P during the luteal phase, it is a more meaningful test than the single P. Therefore, we believe that it can potentially be used in clinical research and in monitoring therapy for ovulatory dysfunction with reasonable assurance. REFERENCES 1. Tredway DR, Mishell DR, Moyer DL: Correlation of endo-
2. 3.
4.
5.
6.
7. 8.
metrial dating with luteinizing hormone peak. Am J Obstet Gyneco1117:1030, 1973 Novak E: Comment. Obstet Gynecol Surv 5:564, 1950 Shepard MK, Senturia YD: Comparison of serum progesterone and endometrial biopsy for confirmation of ovulation and evaluation of luteal function. Fertil Steril 28:541, 1977 Radwanska E, Hammond J, Smith P: Single midluteal progesterone assay in the management of ovulatory infertility. J Reprod Med 26:85, 1981 Abraham GE, Maroulis GB, Marshall JR: Evaluation of ovulation and corpus luteum function using measurements of plasma progesterone. Obstet Gynecol 44:523, 1984 Hammond MG, Talbert LM: Clomiphene citrate therapy of infertile women with low luteal phase progesterone levels. Obstet Gynecol 59:275, 1982 Noyes RW, Hertig AT, Rock J: Dating the endometrial biopsy. Fertil Steril1:3, 1950 Fukushima T, Tajima C, Fukuma K, Maeyama M: Tamoxifen in the treatment of infertility associated with luteal phase deficiency. Fertil Steril 37:755, 1982
Fertility and Sterility
FERTILITY AND STERILITY Copyright c 1987 The American Fertility Society
The integrated luteal progesterone: an assessment of luteal function
Chung H. Wu, M.D.* Shahab S. Minassian, M.D.t Jefferson Medical College of Thomas Jefferson University, and The Medical College of Pennsylvania, Philadelphia, Pennsylvania
An integrated luteal progesterone (ILP L ) was calculated on the basis of a luteal progesterone (P) level with the assumption that the daily plasma P level in the luteal phase closely approximates a sine curve. The midluteal P-amplitude (K) was also obtained mathematically. Daily luteal P levels from five normal ovulatory cycles were assessed for the biologic variation of ILP L and K, then compiled to construct a normogram of the ILP during the luteal phase. The coefficient of variation of K and total ILP L in each cycle ranged from 9.7% to 24.3% and 3.5% to 13.2%, respectively. Fifty-two infertility patients were evaluated for their luteal function by the luteal P and estradiol (E 2) level, K, ILPL , endometrial biopsy (EBX)-lag-day, as well as the lengths of follicular phase, luteal phase (L#), and cycle. Thirty-nine patients had EBX-lag day:$; 2 days and were designated as infertile-normal (INF-NL) luteal phase, while the remaining 13 patients who had EBXlag day> 2 days were considered as luteal phase defect (LPD). Significant (P < 0.05) differences were observed between INF-NL and LPD in: luteal length (13.2 ± 0.31 versus 11.0 ± 0.58 days, respectively), and total ILP L (170 ± 8.3 versus 113 ± 8.5 ng/ml-day, respectively). No differences were seen in luteal P, E2 and K levels, nor in follicular and cycle length. Significant (P < 0.05) correlations were observed between total ILP L and luteal P, E 2, L#, and K; while a negative correlation was noted between follicular and luteal length. Briefly, the total ILPL is a convenient and reliable parameter to assess luteal function in addition to or in place of EBX; and it can be used in clinical research or in monitoring therapy for ovulatory dysfunction. Fertil Steri148:937, 1987
The adequacy of luteal function traditionally has been evaluated most often by the late luteal phase endometrial biopsy (EBX) and, less frequently, by progesterone (P) levels in the serum or plasma and/or luteal length by basal body temperature (BBT). While the EBX is the most widely accepted criterion for luteal phase defect (LPD), it is not without limitations. Criteria for normal versus abnormal readings differ/ large interpersonal varia-
Received March 3, 1987; revised and accepted June 17, 1987. H. Wu, M.D., Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, Thomas Jefferson University, 1025 Walnut Street, Philadelphia, Pennsylvania 19107. t Department of Obstetrics and Gynecology, The Medical College of Pennsylvania.
* Reprint requests: Chung
Vol. 48, No.6, December 1987
tions have been reported, and its accuracy has been questioned. 2 ,3 The expense and discomfort of multiple EBX for diagnosis and monitoring of therapy for LPD is yet another problem. P determinations most accurately reflect luteal P secretion only if clinically impractical daily samples are obtained, and the use of one or three midluteallevels is controversia1. 4 ,5 BBT has been shown to be of some use in the diagnosis and treatment of LPD, 6 but cannot be used alone in this regard. There is a great need therefore for a reliable parameter to assess luteal function in the clinical and clinical research setting. The typical pattern of daily luteal phase P level in the nonpregnant state shows a distinct similarity to a sine curve. The area under the curve of individual cycle was integrated based on the luteal P level Wu and Minassian Integrated luteal progesterone
937
and the luteal length and was designated as the "integrated luteal progesterone" (ILPd. The ILPL then was applied to evaluate the EBX-proven normal and luteal phase defective cycles to assess its usefulness in the clinical evaluation of luteal function.
at exact midpoint of the sine curve. Assigning the luteal length as "L" and the day of blood sampling as "D", then: P = K sin (1rD/L) ILPD =
i i D
P=
D
K sin (1rD/L)
MATERIALS AND METHODS
=
when D
Normal Ovulatory Cycles
Five normally menstruating women with 27- to 30-day intervals were recruited after the proven normalcy of the basic endocrine study and physical examination. The daily blood samples were obtained and analyzed for luteinizing hormone (LH), P, and estradiol (E 2). The data were synchronized at LH peak as day o. The BBT also was obtained as an additional reference to assess the luteal phase length. The end of cycles was designated as the beginning of the next menstruation. Mathematic Formulae for Integrated Luteal Progesterone and Progesterone-Amplitude Level
The P-amplitude level (K) is the largest P level calculated on the sine curve (Fig. 1), which occurs 18
16
r:zzl ILPD o ILPt.
t DAY
Figure 1 The mean (±SD) luteal progesterone (e e) levels from five normal cycles superimposed on a sine curve (0 0). K, P-amplitude; ILPD , integrated luteal progesterone up to the day uD"; ILP L , total integrated luteal progesterone. Formulae: for calculation of K, ILP D and ILPL from P level and the luteal length based on the formula of sine curve.
938
i.e., K = P/sin (1rD/L)
Wu and Minassian Integrated luteal progesterone
=
[1 - cos (1I"D/L)]KLhr
L, ILP L
=
2KL/1I"
where ILPD is the area under the sine curve up to the day "D" and ILPL is the total area under the sine curve that represents the total luteal P secretion. Using these formulae, the K, ILP D , and ILP L were calculated from the daily P data in the luteal phase of five normal cycles. The data from the first and last two days of the luteal phase were excluded from evaluation because they tended to produce higher error. The variation of K and ILPL so-obtained were assessed by the coefficient of variation (CV) (i.e., CV = standard deviation/mean). Study Patients
The luteal phases of 52 infertile women were evaluated during an infertility workup by BBT, EBX, midluteal P, and E2 levels drawn in the morning. Luteal length was determined by BBT, starting on the day when temperature rose above the midpoint between the mean levels of the follicular and luteal phase temperature. EBX were obtained during the late luteal phases of these cycles and dated by the criteria of Noyes et aU Single luteal blood samples were obtained during the middle two thirds of the luteal phase and were measured for P and E2 by standard radioimmunoassay techniques. A diagnosis of a LPD was made when a greater than 2-day lag was noted. Those cycles in which a less than or equal to 2-day lag was seen were designated as normal infertile (INF-NL) cycles. It must be noted that the diagnosis of a single luteal phase defective cycle does not mean that the diagnosis was made of LPD itself as a cause for the infertility. Statistical analysis was performed using Student's t-test. Correlation analysis was obtained by calculating correlation coefficients (r) between paired parameters in all study patients. RESULTS
The mean daily plasma P levels from the luteal phases of five normal cycles are shown in Figure 1. Fertility and Sterility
These women had regular menses, no clinical signs of androgen excess, and normal endocrine profiles. As shown in Figure 1, the daily luteal P level approximates a sine curve and is superimposed on an actual sine curve. The calculated mean values of the K and ILPL from these cycles ranged as follows: K, 14.4 to 20.5 with total mean 16.4 ng/ml; and ILPL , 128 to 185 with total mean 147 ng/ml/day. The coefficients of variation of K and ILP L for individual cycles ranged from 9.7% to 24.3% and 3.5% to 13.2%, respectively, and increased to 26.2% and 25.8%, respectively, when all cycles were combined. The mean daily cumulative ILP n for these cycles is shown in Figure 2. The shaded area includes the mean ILP n ± 1 standard deviation (SD). The portion of the graph below the shaded area represents inadequate P secretion, while the area above represents adequate secretion. Of the infertile patient cycles, 39 cycles were considered normal, and 13 cycles were diagnosed as LPD by EBX. The calculated ILPL and K, along with the midluteal P and E2 levels, and the length of follicular phase, luteal phase, and cycle are shown in Table 1. The mean ILPL of LPD cycles was significantly (P < 0.01) lower than in INF-NL cycles. There was no difference, however, in mean P level between groups. When cycles in which P levels were drawn on the third or fourth day prior to menses were analyzed, the ILP L of LPD cycles were again significantly (P < 0.01) lower than INF-NL. The luteal phase length of LPD cycles
110
110
Table 1 The Results of Ovarian Function Assessment in the Infertile-Normal (INF-NL) and the Luteal Phase Defect (LPD) Cycles Parameter
Unit
INF-NL (n = 39)
LPD (n = 13)
Midluteal P Midluteal E2
ng/ml pg/ml
14.6 ± Loa 157 ± 19
EBX lag-day
day
0.35 ± 0.11
3.1 ± 0.39 b
Follicular length Luteal length Cycle length
day day day
14.6 ± 0.44 13.2 ± 0.31 27.5 ± 0.36
15.5 ± 0.77 11.0 ± 0.58 b 26.5 ± 0.71
P-amplitude (K) ILPL
ng/ml ng/ml/day
19.8 ± 1.3 170 ± 8.3
18.0 ± 1.6 113 ± 8.5 b
10.5 ± 2.0 128 ± 11
Samples drawn on the 3rd or 4th day before menses
Midluteal P ILPL a b
ng/ml ng/ml/day
(n = 23)
(n = 8)
13.6 ± 1.2 171 ± 11.5
9.2 ± 2.8 104 ± 7.6 b
Mean ± SE. P < 0.01 when compared with the respective INF-NL value.
were significantly (P < 0.01) shorter than INF-NL cycles (11.0 ± 0.58 day versus 13.2 ± 0.31, respectively), but no difference in follicular phase length or cycle length was seen. When an ILPL cut-off of 130 ng/ml-day was used to evaluate the study patients, 11 of the 22 (50%) cycles below this level were false-positive as they were normal by EBX. Conversely, 2 of 30 (7%) cycles above this level were false-negative as they were LPD by EBX (sensitivity of93%). An analysis of levels taken on the third and fourth days prior to menses revealed essentially the same results. Correlation analysis was performed on the parameters assessed for ovarian function in the study patients. ILPL was significantly (P < 0.05) correlated with midluteal P (r = 0.662), E2 (r = 0.399), K (r = 0.875), and luteal length (r = 0.495); while K also was correlated with midluteal P (r = 0.681) and E2 (r = 0.401). Additional correlations were noted between cycle length and midluteal E2 (r = 0.422), follicular length (r = 0.736), and luteal length (r = 0.304); and midluteal E2 versus midluteal P (r = 0.439). A significant negative correlation was only observed between follicular length and luteal length (r = -0.532).
20
o
DISCUSSION
o
I.
~
Q
M 00
WTtAI. PHASE
Figure 2 The mean (±SD) cumulative integrated luteal progesterone (ILPD ) in the luteal phase of five normal cycles. Vol. 48, No.6, December 1987
Serum P levels have long been used as an indicator of normal and abnormal luteal function in the infertile woman; however, there has been much Wu and Minassian Integrated luteal progesterone
939
controversy over their value. Daily P levels taken throughout the entire luteal phase would be an ideal method but, obviously, is clinically impractical. Abraham et al5 reported on the use of three midluteal levels with good results, but this is an expensive and inconvenient option. Radwanska et al4 and Hammond and Talbert6 have advocated the use of a single midluteal P level, reporting good results. The advantages of less inconvenience, less cost, and clinical practicality are certainly clear, however, because of reported variations in the single level, the method has gained little general acceptance and the EBX remains the standard for LPD diagnosis. Although there is no denying the usefulness of the EBX as a luteal phase parameter, questions concerning its accuracy, high variation, the patient discomfort, and expense of multiple biopsies for the diagnosis and stepwise evaluation of therapy limits the use of this method in infertility therapy and clinical research. The integration of P levels during the luteal phase has been reported previously,S but no studies have yet proven the adequacy or inadequacy of these cycles by EBX or reported on the exact method of integration. Moreover, it relied on cumbersome daily P determinations. The secretion of P during the luteal phase approximates a sine curve, as evidenced by the five normal cycles in this study. By integrating and representing the entire curve of corpus luteum P secretion with one sample, the ILP L combines the reliability of daily measurements with the convenience of the single level. Also, by virtue of the great difference in the mean ILP L between biopsy-proven LPD and normal cycles, the ILPL appears to reflect what would be found on biopsies and may be used as an adjunct or even a replacement for the EBX when multiple biopsies are indicated. The ILPL also proves to be more reliable than single P levels because no difference in the mean P levels between groups was found in this study. Regardless of the convenience of the single P determination (for ILPd versus multiple levels, a separate vi~it to the laboratory or office would theoretically be required, and would mean two luteal phase visits if an EBX were performed in that cycle. An ideal clinical situation, therefore, would consist of a simultaneous EBX and P taken during
940
Wu and Minassian Integrated luteal progesterone
one visit. Because the EBX is accurate when performed during the late luteal phase, the ILPL results from this segment were examined. It was found that LPD cycles had a much lower mean ILPL than INF-NL cycles, suggesting that only one well-timed visit is necessary. To use the ILPL in an individual patient in the clinical setting, it would be helpful to establish a cutoff for normal levels. When the 130 ng/ml-day level was used for this purpose, 93% of normal ILPL agreed with EBX results. Conversely, nearly 50% of ILPL below 130 ng/ml-day were from INF-NL cycles. Should the clinician use this cutoff, therefore, it appears that, for individual cycles, a normal ILPL may be more meaningful than an abnormal one, and that a low ILPL level during infertility therapy may necessitate a confirmatory EBX. For research purposes, the mean ILPL does show quite a significant difference between groups, and can be used with assurance in this regard. In summary, the ILP L is a convenient, reliable test for luteal phase adequacy. By reflecting the entire secretion of P during the luteal phase, it is a more meaningful test than the single P. Therefore, we believe that it can potentially be used in clinical research and in monitoring therapy for ovulatory dysfunction with reasonable assurance. REFERENCES 1. Tredway DR, Mishell DR, Moyer DL: Correlation of endo-
2. 3.
4.
5.
6.
7. 8.
metrial dating with luteinizing hormone peak. Am J Obstet Gyneco1117:1030, 1973 Novak E: Comment. Obstet Gynecol Surv 5:564, 1950 Shepard MK, Senturia YD: Comparison of serum progesterone and endometrial biopsy for confirmation of ovulation and evaluation of luteal function. Fertil Steril 28:541, 1977 Radwanska E, Hammond J, Smith P: Single midluteal progesterone assay in the management of ovulatory infertility. J Reprod Med 26:85, 1981 Abraham GE, Maroulis GB, Marshall JR: Evaluation of ovulation and corpus luteum function using measurements of plasma progesterone. Obstet Gynecol 44:523, 1984 Hammond MG, Talbert LM: Clomiphene citrate therapy of infertile women with low luteal phase progesterone levels. Obstet Gynecol 59:275, 1982 Noyes RW, Hertig AT, Rock J: Dating the endometrial biopsy. Fertil Steril1:3, 1950 Fukushima T, Tajima C, Fukuma K, Maeyama M: Tamoxifen in the treatment of infertility associated with luteal phase deficiency. Fertil Steril 37:755, 1982
Fertility and Sterility