Peritoneal macrophages and infertility: the association between cell number and pelvic pathology

FERTILITY AND STERILITY Copyright 1985 The American Fertility Society Vol. 44, No. 6, December 1985 Printed in U.S.A. Peritoneal macrophages and infertility: the association between cell number and pelvic pathology David L. Olive, M.D. *t J. Brice Weinberg, M.D.:j: Arthur F. Haney, M.D.* Duke University Medical Center and Veterans Administration Medical Center, Durham, North Carolina Increased numbers of peritoneal macrophages have been repeatedly associated with infertility. Because the factors contributing to this intraperitoneal exudate are unknown, this study was carried out to determine which anatomic or endocrinologic abnormalities in infertile women might be associated with an increase in leukocyte numbers. The peritoneal fluid from 103 women was analyzed. Nonparametric data analysis demonstrated significantly greater cell counts in infertile women with endometriosis, compared with other infertile women (P < 0.01) or fertile control subjects (P < 0.005). Multiple regression analysis was then used to determine the relationship of individual variables to cell number without the influence of confounding factors. These data demonstrate that the best correlation with elevated macrophage number is in women who have infertility and no mechanical fertility factors (of which mild endometriosis is a subgroup). Thus, an increase in peritoneal macrophage number is not restricted to women with endometriosis but, rather, is seen in a subset of infertile women generally without mechanical or endocrinologic infertility factors. Fertil Steril 44:772, 1985 Increased numbers of peritoneal macrophages have been repeatedly associated with infertility. 1-3 Phagocyte-mediated cytotoxicity has been demonstrated for sperm, 4 oocytes, 5 and preimplantation embryos. 6 Furthermore, a localized inflammatory exudate with its attendant leukocyte cytotoxicity has been strongly implicated as the mechanism of action of intrauterine con- Received June 12, 1985; revised and accepted August 21, 1985. *Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology, Duke University Medical Center. tpresent address and reprint requests: David L. Olive, M.D., University of Arkansas Medical School, 4301 West Markham, Slot #518, Little Rock, Arkansas 72205. *Veterans Administration Medical Center and Division of Hematology/Oncology, Department of Medicine, Duke University Medical Center. 772 Olive et al. Peritoneal macrophages and infertility traceptive devices. 7-9 These facts together have led us to postulate that an intraperitoneal inflammatory process adversely influences fertility. The factors contributing to this intraperitoneal inflammation are unknown. Several previous studies noted increased numbers of macrophages in association with endometriosis. 1-3 However, these investigations used heterogeneous control groups and did not correct for potentially confounding variables. As a consequence, the precise relationship between infertility, peritoneal macrophage number, and pelvic pathology has not been well defined. This study was undertaken to delineate those factors that are associated with elevated numbers of peritoneal macrophages. A variety of anatomic and endocrinologic abnormalities are evaluated to identify the role of each in the genesis of intraperitoneal inflammation in infertile women. Fertility and Sterility

Table 1. Frequency of Occurrence of Variables Variable No. % Fertile 19 18.4 Infertile 78 75.7 lndeterminatea 6 5.8 Adhesions 37 35.9 No adhesions 66 64.1 Tubal obstruction 23 22.3 Patent tubes 80 77.7 Ovulatory dysfunction 6 5.8 Normal ovulation 97 94.2 Endometriosis 42 40.8 No endometriosis 61 59.2 Leiomyomata 15 14.6 No leiomyomata 88 85.4 ~hese patients are excluded for the remainder of the analysis. MATERIALS AND METHODS One hundred three women who underwent laparoscopy or laparotomy at Duke University Medical Center comprised the population of this study. The couples had been unable to conceive for at least 1 year. Patients with a history of salpingooophoritis had been free of all signs and symptoms of active infection for at least 1 year. All experiments were conducted with the approval of the Duke University Clinical Investigation Committee, and all patients gave informed consent. At the time of operation, all patients were classified as to the presence or absence of the following: (1) endometriosis, (2) leiomyomata, (3) pelvic adhesions significantly distorting normal pelvic anatomic relationships, ( 4) bilateral tubal obstruction, and (5) ovulatory dysfunction. If endometriosis was present, patients were classified according to the revised classification system of The American Fertility Society (R-AFS). 10 Infertile couples having a male factor (as judged by semen analysis), poor postcoital tests, or significant uterine anomalies were excluded from analysis. Peritoneal fluid (PF) was obtained during laparoscopy as previously describedy Briefly, the fluid was aspirated from the anterior uterovesical space and posterior cul-de-sac through a hollow cannula under direct vision; all of the free PF was removed. The fluid was anticoagulated with 10 U sodium heparin/ml PF and chilled on ice. Ten to 20 J.Ll of PF was diluted 1 to 10 with Turk's blooddiluting fluid (Ricca Chemical Company, Ar- lington, TX), and the white cells were counted. The volume ofpf was then measured to the nearest 0.1 ml and the number ofmacrophages determined. Differential cell counts revealed ~ 90% macrophages consistently, as noted in previous studies. 1 12 Because of the skewed distribution of the data collected, nonparametric statistical analysis was believed to be mandatory. Statistical analysis for pairwise comparisons of groups (i.e., infertile women with endometriosis versus fertile control subjects, etc.) was carried out with the use of the Wilcoxon two-sample test with continuity correction. 13 To determine the relative influence of a single factor upon fluid volume and cell number unencumbered by the effects of other potentially confounding factors, multivariate analysis was performed by the maximum R 2 improvement technique for stepwise regression analysis of multiple independent variables. 14 Correlation coefficients for linear regression analysis between PF volume and total cell number, as well as for comparison of R-AFS point totals to fluid volume and cell number, were performed nonparametrically with the use of the Spearman rank-order test. RESULTS Of the 103 patients evaluated, 78 were infertile at the time of laparoscopy, 19 were parous controls currently undergoing tubal ligation or reanastomosis, and 6 had not attempted conception, so their fertility potential was unknown. The rate of occurrence of each of the variables studied is seen in Table 1. The 42 patients with endometriosis were classified by the R-AFS 10 system as having minimal disease (26 patients, 61.9%); mild disease (8 patients, 19.0%); moderate endometriosis (7 patients, 16. 7%); and severe disease (1 patient, 19.0%) (Fig. 1). Initially, analysis was done on three groups: I, infertile with endometriosis; II, infertile without endometriosis; and III, fertile control subjects. The analysis of the PF volume and total cell number can be seen in Table 2. No significant differences were demonstrated among the three groups when PF volume was evaluated. PF cell number, however, was significantly greater in infertile endometriosis patients than in other infertile women (P = 0.007) or fertile control subjects (P = 0.0013). Vol. 44, No.6, December 1985 Olive et al. Peritoneal macrophages and infertility 773

10 8 0 lljjld 0 0:111,p I 0 0 I MINIMAL 5 0,~8 MILD MODERATE SEVERE 10 15 20 25 30 l8 57 Endometriosis Point Total Figure 1. Endometriosis patient distribution by pomt totals and.stag~ of disease. The women were scored with the R-AFS classification system with minimal, 1 to 5 points; mild, 6. to 15 points; moderate, 16 to 40 points; and severe, > 40 pomts. Therefore, given this association, an attempt was made to determine which findings in the infertile patient might be responsible for the increase in PF cell number. With the use of the infertile patients exclusively (n = 78), multiple regression analysis was carried out for determination of the individual contribution by each of the five variables. Analysis of fluid volume showed that none of the factors investigated seemed to be correlated with an increased volume (Table 3). However, the absence of adhesions (P = 0.002) and ovulatory dysfunction (P = 0.048) did correlate with an increased amount of fluid; i.e., an infertile woman was more likely to have fluid elevation if she did not have significant adhesions or ovulatory dysfunction. Similarly, the absence of tubal obstruction nearly correlated with significant fluid increase (P = 0.084). The presence or absence of endometriosis or leiomyomata, however had no demonstrable association with fluid ' volume. The analysis of total cell number also revealed that none of the factors tested correlated with an increased count. Patients without adhesions, however, did have a higher cell number than Table 2. Characteristics of PF Patient group No. Volume" Cell no." ml X 10 6 I. Infertile with en do- 37 16.1 ± 2.6 22.1 ± 4.2 metriosis II. Infertile without en- 41 12.6 ± 1.5 10.5 ± 1.9 dometriosis III. Fertile control sub- 19 10.5 ± 1.8 7.2 ± 2.5 jects amean ± standard error. those with adhesive disease (P = 0.009) (Table 4). Based on these findings, it was of interest to single out those patients who had no confounding variables in order to view the data in "pure" populations. For patients with endometriosis and leiomyomata as well as those patients with no identifiable infertility factors, the results can be seen in Table 5. Each of these groups has high fluid volumes and cell numbers, with the total generated being the highest yet identified for a subpopulation of infertile women. It thus appears that the best correlation with increased macrophage number and fluid volume is not a unique pathologic diagnosis but rather the presence of infertility without a mechanical cause. Among patients with implants of endometriosis, the relationship between point total by the R-AFS classification system 10 and total fluid volume or cell number was calculated. The Spearman correlation coefficient was significantly negative for R-AFS score, versus total volume (r = - 0.42, P = 0.04); and R-AFS score, versus total cell number (r = - 0.46, P = 0.02). Thus, lower R-AFS scores were associated with greater amounts of fluid and higher cell counts, demonstrating that the extent of disease does not contribute to a higher PF cell number. Finally, the relationship between fluid volume and cell number among all patients can be seen in Figure 2. The correlation coefficient for this distribution is 0.62, differing significantly from 0 (P < 0.01). Thus, to a significant degree the amount of PF reflects the cell count. DISCUSSION High numbers of peritoneal macrophages have been noted repeatedly in infertile women. Furthermore, the peritoneal macrophage number has a direct relationship to the number of intratubal macrophages. 15 Adverse effects of leukocytes upon the reproductive process are well known to the reproductive biologist. Cell-mediated cytotoxicity toward human spermatozoa has been demonstrated in the female reproductive tract both in vivo 4 and in vitrop 16 Similarly, oocyte cytotoxicity has been demonstrated in vitro in an animal model. 5 Animal 6 7 and human 7 8 studies have noted an intrauterine cellular inflammatory response secondary to intrauterine contraceptive devices, and recent evidence suggests leukocytemediated damage to the preimplantation embryo 774 Olive et al. Peritoneal macrophages and infertility Fertility and Sterility

Table 3. Multiple Regression Analysis for Total PF Volume Among Infertile Women Paired variables B value" F p Adhesions versus no adhe- -9.2 10.03 0.002 sions Ovulatory dysfunction ver- -10.6 4.05 0.048 sus normal ovulation Tubal obstruction versus pa- -6.0 3.06 0.084 tent tubes Leiomyomata versus no +3.9 1.09 0.299 leiomyomata Endometriosis implants ver- -2.1 0.44 0.510 sus no endometriosis implants athe B values indicates with which member of the pair a significant correlation is associated (if one exists). A positive value indicates increased volume with the first variable of the pair; a negative value indicates a correlation with the second variable. in the intrauterine-device-bearing monkey. 6 Increased levels of proteolytic enzymes have also been demonstrated in the PF of infertile women.17 These findings together suggest that an inflammatory reaction is present in the peritoneal cavity of a subgroup of infertile women. Despite the identification of an intraperitoneal inflammatory exudate in infertile women, it remains to be determined what factors govern this phenomenon. Three previous studies have addressed the relationship between PF cell number and endometriosis, and all found increased numbers of PF macrophages. In the first report, infertile women with endometriosis were compared with a group of infertile women without endometriosis and a group of fertile control subjects. 1 Halme and associates 3 used similar comparison Table 4. Multiple Regression Analysis for Total Cell Number Among Infertile Women Paired variables B value" F p Adhesions versus no adhe- -126.4 7.28 0.009 sions Tubal obstruction versus pa- -73.3 1.75 0.190 tent tubes Ovulatory dysfunction ver- -95.4 1.28 0.262 sus normal ovulation Endometriosis implants ver- +44.3 0.75 0.389 sus no endometriosis implants Leiomyomata versus no +29.8 0.25 0.621 leiomyomata athe B values indicates with which member of the pair a significant correlation is associated (if one exists). A positive value indicates increasing cell number with the first variable of the pair; a negative value indicates a correlation with the second variable. Table 5. Characteristics of PF in Unexplained Infertile Populations Patient group No. Volume" Cell no." ml X 106 Unexplained infertility 4 22.2 ± 5.4 20.4 ± 9.7 Endometriosis implants 21 22.1 ± 3.9 29.9 ± 6.7 only Leiomyomata only 4 25.9 ± 6.9 18.7 ± 9.2 Total 29 22.6 ± 3.0 27.0 ± 5.1 amean ± standard error. groups with nearly identical findings, although parametric statistics were used in an obviously skewed sample. Finally, Badawy and colleagues 2 compared infertile endometriosis patients with a mixed control group of other infertile patients and fertile sterilization candidates. Their results are in agreement with the other studies, despite the use of parametric statistical testing in a nonnormal distribution. Although these studies demonstrate that elevated PF macrophage counts are associated with infertility and endometriosis, they cannot address the cause of the inflammatory process. We have attempted to determine in this investigation which anatomic or endocrinologic findings were independently associated with high PF leukocyte... 120 110-100 0 :90 iso E ~ 70 ~ 60 a ~50 0 ~ 40 11. ~ 30-0 ~ 20 -............. 10 -. - 0---:-. 0 10 20 30 40 50 60 Total Peritoneal Fluid Volume (ml) Figure 2 Total PF volume versus total number ofperitonealleukocytes. Nonparametric linear regression analysis revealed the distribution to have a significant correlation (r = 0.62, P < 0.01). Vol. 44, No.6, December 1985 Olive et al. Peritoneal macrophages and infertility 775

14 12 :;;10 "" ~8 z 'Es ~ CJ C...4 2 o~~~~~~~~~~~~~~~~- Figure 3 Distribution of total peritoneal leukocyte counts in 29 women with no mechanical or endocrinologic cause for infertility: unexplained infertility, endometriosis implants alone, and leiomyomata only. counts. To test the association of these factors, the use of muitiple regression analysis was felt most appropriate. This technique mathematically normalizes the effect of any confounding variabies so that the true effect of the presence or absence of a single factor can be evaluated. In performing this analysis in infertile women, the presence or absence of endometriotic implants, leiomyomata, tubal obstruction, or ovulatory dysfunction did not appear to influence the peritoneal leukocyte count. The only significant relationship was a negative one, i.e., the presence of a mechanical problem (i.e., adhesions) was associated with lower macrophage counts. Similar findings were noted with PF volume. Based on these data, the subset of infertile women with high intraperitoneal macrophage counts is that group without identifiable mechanical or endocrinologic causes of infertility. Thus, despite the absence of a direct relationship between endometriosis and high macrophage counts, women with milder degrees of endometriosis would be expected to segregate disproportionally more often into that subset. Supporting this is the finding that significantly higher macrophage counts were observed in women with low R-AFS endometriosis scores and no anatomic distortion, compared with those with high scores and presumably considerable anatomic distortion. Given this concept, it is understandable why previous studies, having included patients primarily with mild disease, repeatedly demonstrated a strong association between endometriosis and high intraperitoneal cell counts. Combining all patients without mechanical or endocrinologic causes for infertility, a sub- population was produced with the highest aggregate peritoneal cell number of any report to date (Table 5). Even still, a review of these 29 women with unexplained infertility reveals a bimodal distribution, with roughly half having highly elevated macrophage levels, whereas half had normal cell counts (Fig. 3). Thus, although leukocyte-mediated cytotoxicity might be responsible for some previously unexplained infertility, it does not account for all such patients. Although the occasional patient may have two coexisting problems, the number of high-leukocyte women should be fairly small in the groups of infertile women with a clearly identifiable mechanical or endocrinologic explanation for their low fecundity. This is precisely what is found in this study, with the fewest number of patients with high PF macrophage numbers being seen in women with pelvic adhesions and resultant anatomic distortion. Thus, the previously demonstrated association between endometriosis and elevated peritoneal macrophage number is a complex relationship, and the pathogenesis of these two findings may be governed by independent factors. The common denominator that may link these two pathologic entities is retrograde menstruation. This phenomenon may well be near universal,18 19 and the ability of the endometrial cell to ectopically implant may be related to decreased lymphocytotoxicity directed toward endometrial antigens.20 21 It may well be that intraperitoneal inflammation and leukocytosis is a response by some women to the process (retrograde menstruation), rather than the occasional consequence of that process (endometriosis implants).22 23 Conversely, the possibility exists that this finding is totally unrelated to the development of endometriosis and may represent an associated but independent entity. Further investigation into the origin and effects of elevated numbers of peritoneal macrophages is clearly needed to fully understand its relationship to infertility. In summary, this study extends our previous work that demonstrates an association between macrophage-mediated cytotoxicity and infertility. We have shown that elevated numbers of macrophages are not exclusive to women with endometriosis but rather correlate best with infertility in the absence of mechanical or endocrinologic disorders. This suggests that women with high peritoneal leukocyte counts represent a dis- 776 Olive et al. Peritoneal macrophages and infertility Fertility and Sterility

tinct group of infertile patients in which women with minimal or mild endometriosis are a subset. Acknowledgments. We would like to thank Kerry L. Lee, Ph.D. for his time and assistance in the statistical calculations needed, and Ms. Marcia Hobbs and Ms. Mary Misukonis for their able assistance in the laboratory. REFERENCES 1. Haney AF, Muscato JJ, Weinberg JB: Peritoneal fluid cell populations in infertility patients. Fertil Steril 35:696, 1981 2. Badawy SZA, Cuenca V, Marshall L, Munchback R, Rinas AC, Coble DA: Cellular components in peritoneal fluid in infertile patients with and without endometriosis. Fertil Steril 42:704, 1984 3. Halme J, BeckerS, Hammond MG, Raj S: Pelvic macrophages in normal and infertile women: the role of patent tubes. Am J Obstet Gynecol 142:890, 1982 4. Austin CR: Fate of spermatozoa in the female genital tract. J Reprod Fertil 1:151, 1960 5. Weissmann G, Finkelstein MC, Csernansky J, Quigley JP, Quinn RS, Techner L, Troll W, Dunham PB: Attack of. sea urchin eggs by dogfish phagocytes: model of phagocyte-mediated cellular cytotoxicity. Proc Natl Acad Sci USA 75:1825, 1978 6. Hurst PR, Jefferies K, Eckstein P, Dawson K, Whealer AG: Leukocytes are consistently associated with degenerating embryos in IUD-bearing rhesus monkeys. Nature 269:331, 1977 7. Joshi SG: Local effects of pharmacologically inert IUDs in rats, baboons, and humans. In Analysis of Intrauterine Contraception, Edited by F Hefnawi, SJ Segal. New York, American Elsevier, 1974, p 339-8. Gawad AHA, Toppozada HK, Sawi ME, Saleh F, El Sahwi S: Study of the uterine environment in association with intrauterine contraceptive devices. Contraception 16:469, 1977 9. El-Habashi M, El-Sahwi S, Gawish S, Osman M: Effects of Lippes loop on sperm recovery from human fallopian tubes. Contraception 22:549, 1980 10. American Fertility Society: Revised American Fertility Society Classification of Endometriosis: 1985. Fertil Steril 43:351, 1985 11. van Furth R, Raeburn JA, van Zwet TL: Characteristics of human mononuclear phagocytes. Blood 54:485, 1979 12. Muscato JJ, Haney AF, Weinberg JB: Sperm phagocytosis by human peritoneal macrophages: a possible cause of infertility in endometriosis. Am J Obstet Gynecol 144: 503, 1982 13. Wilcoxon F: Individual comparisons by ranking methods. Biometrics 1:80, 1945 14. Sail J, SAS Regression Applications, Technical Report A-102. Cary, NC, SAS Institute, 1981 15. Haney AF, Misukonis MA, Weinberg JB: Macrophages and infertility: oviductal macrophages as potential mediators of infertility. Fertil Steril 39:310, 1983 16. London SN, Haney AF, Weinberg JB: Macrophages and infertility: enhancement of human macrophage-mediated sperm killing by antisperm antibodies. Fertil Steril 43:274, 1985 17. Halme J, BeckerS, Hammond MG, Raj MHG, Raj S: Increased activation of pelvic macrophages in infertile women with mild endometriosis. Am J Obstet Gynecol 145:333, 1983 18. Halme J, Hammond MG, Hulka JF, Raj SG, Talbert LM: Retrograde menstruation in healthy women and in patients with endometriosis. Obstet Gynecol 64:151, 1984 19. Blumenkratz MT, Gallagher N, Bashore RA, Tenckhoff H: Retrograde menstruation in women undergoing chronic peritoneal dialysis. Obstet Gynecol 57:667, 1981 20. Dmowski WP, Steele RW, Baker GF: Deficient cellular immunity in endometriosis. Am J Obstet Gynecol 141: 377, 1981 21. Steele RW, Dmowski WP, Marmer DJ: Immunologic aspects of human endometriosis. Am J Reprod Immunol 6:33, 1984 22. Olive DL, Hammond CB: Endometriosis: pathogenesis and mechanisms of infertility. Postgrad Obstet Gynecol 5(9):1, 1985 23. Olive DL, Haney AF: Endometriosis. In Reproductive Failure, Edited by AH DeCherney. New York, Churchill Livingstone. In press Vol. 44, No. 6, December 1985 Olive et al. Peritoneal macrophages and infertility 777