Assessing gonadal function after childhood ovarian surgery

International Statistical Classification of Diseases and Related Health Problems, Ninth Revision, codes were used to identify patients from birth to 20 years of age who were surgically managed for an ovarian condition at an academic health care facility consisting of an adult hospital and a freestanding children's hospital between August 2000 and August 2010. One hundred eighty patients were included in the analysis. A single researcher (AZ) abstracted qualifying records for demographics, age at diagnosis, surgical procedure(s) performed, and pathology. To evaluate the late effects of either unilateral oophorectomy or ovarian salvage in childhood on subsequent gonadal function, we conducted a mail survey of all 180 study patients. A panel of researchers developed an 11-item survey to capture effects on the onset of menarche and puberty, history and symptoms of menstrual irregularity, age at start of menstrual irregularity, need for hormone regulation, and ability to achieve pregnancy ( Fig. 1 ). The revised survey was then distributed to the entire study cohort along with a letter of intent and appropriate assent and consent forms. In cases where patients were younger than 18 years at the time of follow-up, parents or legal guardians were provided the option of completing the survey. The institutional review board approved all aspects of the study (no. 091112). Data from the chart review and survey responses were compiled into a computer database, and descriptive univariate analyses were performed. Continuous data were compared using the nonparametric Wilcoxon rank sum test. Proportions were compared using the χ 2 test or Fisher exact test in instances where cell values were less than 5. Significance was set to P < .05. All statistical tests were performed using STATA 11.2 (2009 StataCorp LP, College Station, TX).

Within our academic center, a total of 180 patients (mean age, 11.5 years; range, 4 days to 20 years) were identified as having had an operation to treat an ovarian disorder during the study period. Eighty-six females had unilateral oophorectomy with or without salpingectomy, and 94 had ovary-sparing procedures, including partial oophorectomy, cystectomy, salpingectomy only, tumor enucleation, detorsion, oophoropexy, or drainage. Median follow-up time was 40.5 months for the oophorectomy group and 29 months for the ovarian salvage group. The survey was mailed to all 180 patients, and 81 completed forms were returned (45% response rate). Characteristics of respondents and nonrespondents only differed by race; those responding were predominantly white (n = 67, 81.5%) ( Table 1 ). All other characteristics were similar between response groups. Mean age of respondents was 138 months (interquartile range, 114-173 months) at diagnosis and 178 months (interquartile range, 147-219 months) at survey. Mean age of nonrespondents was 137 months (interquartile range, 98-191 months) at diagnosis and 181 months (interquartile range, 128-232 months) at survey. Simple cyst, benign tumor, and torsion owing to mass were the most common diagnoses in both groups. Unilateral salpingo-oophorectomy, oophorectomy, and cystectomy were the most common procedures. Of the respondents, 44 had unilateral oophorectomy, and 37 had an ovary-sparing procedure. Tables 2 and 3 compare the 2 groups. As expected, a diagnosis of complex or simple cyst (32.4% and 48.7%, respectively) was notably higher in the group with ovary-sparing procedures ( P = .02), whereas a diagnosis of benign or malignant tumor (54.6% and 9.1%, respectively) was more common in the group having an oophorectomy ( P < .0001). The median age at onset of menarche (12 years) was similar between the 2 groups. No difference in achievement of normal menarche, defined as menarche occurring between ages 10 and 15 years, was observed across surgical groups. Painful menses were significantly more common in females having an ovary-sparing procedure (59.5%) when compared with females having a unilateral oophorectomy (27.3%; P = .003). A history of menstrual irregularity, defined as cycles longer or shorter than 28 to 30 days, abnormally heavy or light bleeding, and/or severe pain with menses, was reported in 70% of the patients having ovary-sparing procedures relative to 50% of the patients having oophorectomies who were postmenarche at time of survey ( P = .181). Of 4 women who had a unilateral oophorectomy in childhood and attempted to become pregnant, 3 reported successful conception and delivery, whereas none in the salvage group had attempted to become pregnant. Twenty-three of the survey respondents who underwent ovary-sparing procedures and 19 of the respondents who underwent oophorectomies were postmenarche at the time of surgery. Fig. 2 compares the menstrual cycles of these 2 surgical groups before and after surgery. The number of patients having regular menstrual cycles before surgery and who continued to be regular after surgery was significantly higher in the oophorectomy group with 7 (70%) of 10 patients noting continued menstrual regularity, whereas only 2 (15%) of 13 patients reported regular cycles in the ovarian salvage group ( P = .013). All patients who reported irregular cycles before surgery remained irregular after surgery.

Although current studies recommend conservative treatment of pediatric ovarian disorders to ensure future fertility, a review of the literature yielded little information directly comparing the late effects of ovary salvage vs oophorectomy on subsequent fertility. Studies in adults suggest an association between type of ovarian surgery and subsequent fertility as measured by achievement of pregnancy. In a series reported by Yinon et al [ 15 ] comparing the results of cystectomy vs oophorectomy for borderline epithelial tumors, the fertility rate was lower in patients treated with the gonadal sparing cystectomy. Perhaps leaving an injured ovary in some way hinders fecundity. On the other hand, in a separate study of borderline tumors comparing unilateral salpingo-oophorectomy and contralateral cystectomy or bilateral cystectomies in patients treated for bilateral tumors, the authors showed a higher cumulative pregnancy rate in patients having undergone bilateral cystectomies compared with unilateral oophorectomy and contralateral cystectomy [ 16 ]. These results suggest that cystectomy should be preferred, if technically feasible. However, this latter trial was limited by a small number of patients (15 in the cystectomy group and 17 in the cystectomy with oophorectomy group). Because achievement of pregnancy is an unrealistic outcome variable in the pediatric population, we felt that an examination of menstrual events would be an appropriate surrogate assessment of gonadal function in females who required ovarian surgery. Although menstrual function may be perceived as an inaccurate indicator of future fertility, early pioneering studies on endometrial dating by Rock and Bartlett [ 17 ] revealed a significant correlation between ovulatory function and regularity of the menstrual interval and flow pattern. A more recent article documents a relationship between 30- to 31-day menstrual cycles and 5-to 7-day bleeding periods and fecundity [ 18 ]. In addition, Gardner [ 19 ] found that women experiencing dysmenorrhea and cycle irregularity in the first few years postmenarche were at risk for having poor gynecological health later in life. Although painful menses with no pelvic pathology and normal ovulatory cycles (primary dysmenorrhea) is not linked to infertility [ 20 ], dysmenorrhea in the context of established ovarian pathology and irregular cycles as well as previous surgery is important to consider when discussing gonadal function and fecundity. Secondary dysmenorrhea is highly correlated with gynecological pathologies that have well-documented effects on fertility [ 21 ]. Disease processes such as endometriosis, ovarian/pelvic adhesions, and large hormone-secreting cysts have been shown to disrupt folliculogenesis, reduce fertilization potential of oocytes as well as inhibit oocyte release and transport [ 22 - 24 ]. In addition, the likelihood of developing these disorders appears to increase with pelvic surgery [ 21 , 25 ], and recurrence rates can be high even after surgical intervention [ 9 , 26 , 27 ]. Menstrual cycle irregularity and secondary dysmenorrhea are likely associated with reduced fertility and impaired gonadal function and are certainly associated with patient discomfort and anxiety. When we compared females with ovary-sparing procedures to females with unilateral oophorectomies, there was a notable increase in menstrual irregularity and painful menses in the ovarian salvage group. No difference between these groups was observed in achievement of normal menarche or age at onset of menarche. In our cohort, oophorectomy did not appear to delay or impair long-term hormonal function when compared with ovarian salvage. Lass [ 14 ], who performed a comprehensive literature review of reproductive performance in adult women with a single ovary, also concluded that absence of an ovary does not reduce fertility potential. However, women have no compensatory mechanism for the loss of 1 ovary, and because the number of primordial follicles in the ovary is finite, these women may have a shorter reproductive lifespan [ 14 ]. An alternative explanation for the increase in menstrual irregularity observed in the ovarian salvage group may be differences in initial underlying disease etiology and pathology between the 2 surgical groups. The salvage group had more cystic disease, which is hormonally driven and, hence, presumably more likely to affect menses, as opposed to the oophorectomy group, which predominantly had tumors. However, when we followed the menstrual cycles of females who were postmenarche at the time of surgery from preoperative to postoperative, our data showed similar percentages of irregularity in both groups before surgery and a marked increase in irregularity in the salvage group after surgery. Interestingly, continuation of menstrual regularity from preoperative to postoperative was significantly higher in the oophorectomy group vs the ovary salvage group. If initial underlying pathology were solely responsible for the differences in menstruation seen in the 2 surgical groups, we would predict a higher percentage of irregularity in the salvage group before surgery and similar menstruation patterns across groups once the offending disease had been eradicated. Instead, the opposite trend was observed. Perhaps the presence of an altered ovary somehow disorders menses in perimenarchal women. Although the effects of pathology should not be discounted, oophorectomy may surprisingly increase the chances of preserving normal ovarian activity. Interpretation of our study results is limited by the retrospective study design coupled to a survey. Retrospectively obtained data are limited to information already recorded within the record for purposes of clinical care. However, we attained a respectable 45% response rate to the survey. Notably, those that responded were primarily of white race. Respondents also had oophorectomies in greater proportion. These observations may indicate differential response rates within subsets and could bias our findings. Symptomatic patients may have responded in greater proportion too. These factors may have influenced the differential response rates observed yet, if true, should have increased the amount of respondents reporting painful menses in those who had oophorectomies. Differential response rates did not otherwise impact the survey in substantive ways. Nevertheless, we did not find any significant differences between respondents and nonrespon-dents and so conclude that this subset is representative of the entire cohort. In addition, as with all surveys, it is possible that the language of each question biased a patient's or guardian's response in a manner that could have altered the study findings. This study should be reassuring to patients, guardians, and care providers, having or treating an ovarian condition, which requires unilateral oophorectomy, in that loss of a single ovary does not appear to adversely affect gonadal function, as estimated by effects on menstrual regularity, nor does it appear to compromise future fertility. However, we continue to advocate the conservative approach to spare ovaries whenever possible and feasible. Notably, the sparing of ovaries in this population of young females may result in painful and/or irregular menses, which will be beneficial knowledge for patients, guardians, and care providers. To the best of our knowledge, this study is the first to attempt to address the controversy of late effects of childhood ovarian surgery on subsequent ovarian function. Larger prospective, long-term follow-up studies are needed to shed more light on this important topic.