When two diseases collide: Endometriosis as a risk factor for severe pelvic inflammatory disease

MC was responsible for data collection. MC, MR, and CR contributed to study design, data analysis, and drafting of the manuscript. All authors critically reviewed the manuscript and approved the final version.

The study received approval from the local institutional ethical review board (HCB/2025/0307) on 15 May 2025.

The authors have nothing to report.

Between January 2018 and December 2023, a total of 174 patients were initially included in the study. Of these, 15 were excluded for not meeting the inclusion criteria. Among the remaining 159 patients, 48 (30.2%) were diagnosed with endometriosis, and 111 (69.8%) were classified into the no endometriosis group. Table  1 summarizes the demographic data and baseline characteristics of these patients. The groups were comparable with respect to parity, nationality, previous surgery, and use of hormonal contraception at admission. Of note, women with endometriosis were significantly older compared to those without endometriosis. Regarding the potential trigger of PID by invasive procedures, only one event of oocyte retrieval was reported in an endometriosis patient just prior to the onset of PID. Additionally, one endometriosis patient and two without endometriosis had undergone intrauterine device insertion within the 3 months preceding the PID episode. Furthermore, 30/48 (63%) of women in the endometriosis group were not under hormonal treatment at the time of admission. Demographic data and baseline patient characteristics in patients with pelvic inflammatory disease with and without endometriosis. Note : Data are reported as % ( n ) unless otherwise indicated. Table  2 summarizes the TVUS findings, symptoms, and laboratory results at admission in patients with and without endometriosis. It was of note that patients with endometriosis had significantly higher rates of unilateral or bilateral adnexal findings, including tubo‐ovarian abscess, suspected infected ovarian endometrioma, and pyosalpinx. The most common adnexal finding in patients with endometriosis was ovarian endometrioma. In contrast, pyosalpinx was the most frequent finding in patients without endometriosis, followed by tubo‐ovarian abscess. The groups were similar in terms of symptoms at admission, except for fever, which was more common in patients without endometriosis. They also showed comparable laboratory findings, including C‐reactive protein and leukocyte levels. Presenting symptoms, results of blood analyses and cultures at admission in patients with pelvic inflammatory disease with and without endometriosis. Note : Data are reported as n (%) unless otherwise indicated. The bacteria identified were classified as follows: Sexually Transmitted disease (including Chlamydia, Neisseria gonorroheae, mycoplasma ). Fungi (including Chlamydia albicans ). Gram positive (including Enterococcus faecalis , coagulase‐negative Staphylococcus, Gardnerella vaginalis , Streptococcus anginosus , and Corynebacterium spp.). Gram negative (including Escherichia coli , Bacteroides fragilis , Citrobacter freundii complex , Bacteroides thetaiotaomicron , and Pseudomonas spp). Abbreviations: CRP, C‐Reactive Protein; GI symptoms, Gastrointestinal symptoms; Gram −, Gram negative; Gram +, Gram positive; OMA, Endometrioma; STD, Sexually Transmitted disease; TOA, Tubo‐ovarian abscess. The majority of vaginal and endocervical cultures in patients with endometriosis were negative compared to the results of patients without endometriosis. Vaginal or endocervical culture showed that only one patient with endometriosis tested positive for sexually transmitted disease (STD), whereas 47/111 patients (42.3%) without endometriosis had positive STD cultures ( p = 0.01 ). Only 11 out of 159 patients (6.9%) underwent surgical intervention during hospitalization: 6 in the endometriosis group and 5 in the no endometriosis group. Regarding intraoperative cultures, among patients with endometriosis, two were negative, one tested positive for Streptococcus intermedius , and one for Escherichia coli , while samples were not taken in two cases. Among patients without endometriosis, four cultures were negative, and one tested positive for Streptococcus anginosus . Table  3 shows the clinical outcomes in patients with and without endometriosis. In the unadjusted analysis (Table  3A ), treatment failure and recurrence rates were significantly higher in patients with endometriosis, while the need for surgery was higher, albeit not statistically significant, in patients with endometriosis. Women with endometriosis also required a longer hospital stay. After multivariate adjustment (Table  3B ), endometriosis remained independently associated with treatment failure, recurrence, and prolonged hospital stay, but not with the need for surgery. Unadjusted and multivariate‐adjusted associations between endometriosis and primary outcomes in patients with pelvic inflammatory disease PID. Note : Data are reported as n (%) unless otherwise indicated. Multivariate logistic regression models were used for binary outcomes and linear regression for length of hospital stay. All models were adjusted for age, prior pelvic surgery, and hormonal treatment at admission. Adjusted OR are shown for treatment failure, surgical intervention and recurrence. The β coefficient is reported for length of hospital stay. Abbreviations: aOR, Adjusted Odds Ratio; CI, confidence interval. Within the cohort of patients diagnosed with endometriosis, the subtype of the disease was characterized using TVUS during the second scan performed 1 month after admission. Ovarian endometrioma was the most frequent phenotype (72.9%) and was consistently identified in both examinations. On follow‐up, deep endometriosis was confirmed in 31/48 patients (64.5%), although only 17 were initially suspected at admission. Conversely, adenomyosis was confirmed in 8/48 cases (16.7%), while two patients were overdiagnosed at initial assessment but not confirmed on follow‐up. As shown in Table  4 , clinical outcomes were subsequently analyzed within the endometriosis group, comparing patients with and without deep endometriosis. Although no statistically significant differences were found in any of the outcomes, patients with deep endometriosis showed higher rates of treatment failure and recurrence. The median length of hospital stay was similar between groups; however, patients with deep endometriosis showed greater variability, a wider range, and increased skewness. Comparison of clinical outcomes between patients with pelvic inflammatory disease with and without deep endometriosis. Note : Table  4 comparative analysis of clinical outcomes among patients with PID with and without DE. Data are presented as n (%) for categorical variables and as median (IQR) with range for continuous variables.

In our cohort, PID in patients with endometriosis was independently associated with significantly higher rates of recurrence, increased antibiotic treatment failure, and longer hospital stay compared to those without endometriosis. Moreover, the majority of vaginal cultures in these women were negative. Ovarian endometrioma was found to be the most common subtype of endometriosis, and the presence of deep endometriosis was associated with more severe, although not statistically significant, clinical outcomes. Additionally, we observed a significantly lower prevalence of STDs in patients with endometriosis compared to those without. The prevalence of endometriosis in our cohort was notably higher than that reported in the general population (30.2% vs. 10%), 1 in concordance with other studies. 2 , 3 Of note, other studies, 4 have observed a prevalence of endometriosis of up to 63% among patients undergoing surgery for PID. These findings support a potential predisposition of women with endometriosis to develop PID. Although PID is traditionally associated with younger age (< 25 years) and multiple sex partners 7 —risk factors linked to STD—our findings suggest that patients with endometriosis tend to be older. 2 This older age profile may reflect the influence of endometriosis‐related risk factors, such as comorbidities or pelvic anatomical distortions, that could increase the susceptibility to PID despite the absence of classical risk factors. This is further supported by the fact that STDs were identified in only 2.1% of the cultures of patients with endometriosis. Furthermore, women with endometriosis show alterations in gut and vaginal microbiota, with atypical bacteria such as E. coli contributing to elevated endotoxin levels in menstrual and peritoneal fluid. 10 , 11 , 12 While E. coli is commonly reported in 33–42% of tubo‐ovarian abscesses, 13 , 14 it was detected in only one endometriosis patient in our cohort. We hypothesized that these microbial shifts, along with immune dysfunction and pelvic anatomical changes, may define a distinct pathophysiological pathway for PID in women with endometriosis that diverges from the classical presentation seen in young, sexually active women with STDs. 7 This could explain the differences in the clinical course, risk factors and age at presentation in these patients. It is noteworthy that more than half of the women with endometriosis in our cohort were not receiving hormonal treatment at admission. Hormonal therapy has been shown to reduce menstrual bleeding and modulate inflammatory activity, 15 , 16 factors that might have protective effects against PID. This aspect warrants further investigation as the role of hormonal therapy in PID severity remains unclear. In accordance with previous research, 1 , 17 , 18 we also suggest reconsidering the assumption that severe PID in patients with endometriosis is typically preceded by an invasive procedure. In our cohort, we only identified one case of oocyte retrieval and one case of intrauterine device insertion immediately preceding the onset of PID. With respect to clinical outcomes, hospital stays in patients with endometriosis were almost twice as long compared to those without, in line with previous studies, 1 , 2 suggesting a more prolonged and refractory course in this group. In our multivariate analysis, endometriosis was independently associated with a mean increase of 1.7 days of hospitalization (95% CI: 0.6–2.8, p = 0.01 ). Antibiotic treatment failure was also higher among women with endometriosis, consistent with previous studies, 2 which reported a failure rate of 48% in patients with endometriosis, exceeding the 25% failure rate previously reported in the general population with tubo‐ovarian abscess hospitalized due to PID. 4 , 19 PID recurrence was significantly more frequent in the endometriosis group, occurring nearly one‐third more often than in those without the disease. Endometriosis was found to be an independent risk factor for both treatment failure and recurrence, with a three‐fold increased risk for each outcome. It is noteworthy that only 6.9% of all women required surgical intervention during hospitalization, with a higher surgical rate observed among those with endometriosis, consistent with previous studies. 1 , 3 This increased need for surgical management reflects the greater disease complexity of endometriosis cases, with longer procedures, increased blood loss and extended hospitalization. 4 However, the overall low rate of surgery likely reflects a predominantly conservative approach aimed at reducing acute inflammation and postponing intervention when feasible. This is particularly important in patients with endometriosis, in whom delayed elective surgery by experienced teams is generally preferred. The worse clinical outcomes observed in women with PID and endometriosis may be explained by several underlying mechanisms associated with endometriosis, including elevated levels of pro‐inflammatory cytokines, altered immune responses, and anatomical distortions resulting from fibrosis and adhesions. 20 , 21 , 22 Additionally, the static pelvic environment may favor bacterial persistence and recurrent infections, while chronic inflammation can reduce the effectiveness of antibiotics. 13 Although most studies do not differentiate between specific adnexal pathologies 1 (tubo‐ovarian abscess, ovarian endometrioma, and hydrosalpinx), TVUS evaluation enabled reliable distinction among these diseases (Figure  1 ), which allowed the identification of notable differences between patients with and without endometriosis. In particular, adnexal findings were more prevalent in women with endometriosis, with ovarian endometrioma representing the most frequent subtype. This higher prevalence of ovarian endometriomas may provide a favorable environment for bacterial proliferation due to their hemorrhagic content, potentially increasing the risk of antibiotic failure and the need for surgical intervention. Of note, during the acute phase of PID, the evaluation of deep endometriosis may be particularly challenging due to the presence of inflammatory artifacts such as edema, fluid collections, or fibrin, which can both mimic endometriotic lesions and impair their visualization. Posterior compartment assessment is often limited by patient discomfort and pain. In our cohort, deep endometriosis was suspected in half of patients at admission. In contrast, adenomyosis appeared to be overdiagnosed during the acute phase, likely reflecting transient inflammatory changes. To avoid misclassification, we systematically performed a follow‐up TVUS, and the diagnosis of endometriosis was only confirmed when findings persisted in this second scan. The potential impact of deep endometriosis on the clinical course of PID along with these specific sonographic features during acute inflammation represents unexplored areas that deserve future research. Our PID patients with associated deep endometriosis showed higher rates of treatment failure and recurrence compared to those without deep endometriosis, although none of these differences reached statistical significance (Table  4 ). Notably, all patients with deep endometriosis exhibited adnexal findings on imaging study. These observations suggest a potentially more unfavorable clinical course in this group, a relationship that has not been previously explored. The present study has notable strengths. It includes a large cohort of patients, which increases statistical robustness; however, its retrospective and single‐center design inherently introduces some methodological limitations. Nevertheless, being a single‐center study and referral hospital for endometriosis ensures consistency in clinical management and diagnostic standards but may limit the generalizability of our findings. Another limitation is that anaerobic cultures are not routinely performed in our institution, which may have led to underestimation of their role in PID. Additionally, although histological confirmation of endometriosis was not available, the diagnosis was based on TVUS performed by experienced sonographers using validated protocols. Given the strong evidence supporting TVUS as a reliable diagnostic tool, this approach is fully aligned with current clinical standards. 23 Also, although endometriosis was confirmed on follow‐up TVUS after clinical resolution of PID, residual inflammatory changes may have persisted in some patients, and therefore a degree of diagnostic misclassification cannot be completely excluded. Furthermore, we acknowledge the potential for surveillance bias, as patients with endometriosis may have been more closely followed in our institution, which is a referral center for this condition. Additionally, as only hospitalized patients were included, the findings of this study may not be generalizable to patients with milder forms of PID managed in the outpatient setting. Lastly, although the comparison between PID patients with and without deep endometriosis was based on small subgroups, to our knowledge, this has not been previously explored and warrants further investigation through adequately powered studies.

Patients with endometriosis and PID are at increased risk of a more severe clinical course. These findings suggest the need to adapt current management protocols. Empirical antibiotic regimens may require modification to cover non‐STD pathogens more effectively, and admission criteria should consider the likelihood of longer hospital stays. In this context, upfront initiation of broad‐spectrum antibiotic therapy could be considered in patients with endometriosis. As these cases are less likely to involve STD‐related PID, clear communication is essential to avoid unnecessary social or emotional distress. TVUS plays a key role in the diagnostic process, helping to identify both adnexal complications and features suggestive of endometriosis, which can guide early and tailored management. Overall, personalized diagnostic and therapeutic approaches, along with clear and reassuring patient communication, are essential to improve outcomes in this high‐risk population. These observations call for future research to better define optimal management strategies.

Recent studies suggest that women with endometriosis are at greater risk of developing pelvic inflammatory disease (PID) 1 and may be predisposed to presenting a more severe course, 2 , 3 leading to worse outcomes such as prolonged hospitalization, increased rates of antibiotic treatment failure, and increased need for surgical intervention. 4 , 5 The association between endometriosis and PID severity may be due to several factors, including altered immune response with increased inflammatory cytokine production, 6 impaired ability to ward off bacteria, and anatomical changes associated with endometriotic lesions, 4 which may facilitate the spread of infection. The diagnosis of PID in patients with endometriosis is complex. Overlapping symptoms, such as pelvic pain and adnexal tenderness, may lead to misdiagnosis or suboptimal management of this multifaceted, fertility‐compromising condition. 4 In addition, imaging interpretation is particularly challenging due to the presence of inflammatory artifacts such as fibrin, which can mimic endometriotic lesions. Furthermore, the heterogeneity of endometriosis subtypes is rarely considered in the literature. Current research on PID in women with endometriosis primarily focuses on ovarian endometrioma, whereas the potential impact of deep endometriosis on clinical outcomes remains unexplored. Therefore, the aim of the study was to evaluate the severity of PID in hospitalized women with and without endometriosis and compare their demographic characteristics, ultrasound findings, and clinical outcomes. Additionally, a subgroup analysis was conducted within the endometriosis cohort to determine whether the presence of deep endometriosis further exacerbates disease severity.

Authors declare no conflict of interest.

A retrospective cohort study was conducted in the Gynecology Department at Hospital Clinic of Barcelona, a tertiary university hospital with a referral unit for endometriosis. We included all women admitted to the hospital with a diagnosis of PID from January 2018 to December 2023. Eligible patients were identified through the hospital electronic medical record system, where PID diagnoses are routinely coded and recorded as a discharge diagnosis. Patients were divided into two groups based on the diagnosis of endometriosis (Group 1: Endometriosis; Group 2: No endometriosis). The diagnosis of PID was made according to the criteria established by the Centers for Disease Control and Prevention (CDC), 7 requiring at least one of three minimum clinical criteria: acute pelvic pain, cervical motion and/or uterine tenderness. Additionally, one of the following criteria had to be present: body temperature > 38.3°C, abnormal cervical mucopurulent discharge or cervical friability, elevated erythrocyte sedimentation rate or C‐reactive protein or laboratory documentation of cervical infection with Neisseria gonorrhoeae or Chlamydia trachomatis . Patients were excluded if an alternative diagnosis was established (e.g., ovarian torsion or appendicitis), or if hospitalization was not required according to the clinical protocol at our institution. The institutional protocol follows the recommendations of the CDC, with hospitalization indicated for patients presenting severe clinical illness, tubo‐ovarian abscess, inability to tolerate oral treatment, lack of response to outpatient therapy, or when a surgical emergency could not be excluded. Due to the retrospective nature of the study, no prior sample size calculation was performed. Endometriosis was initially suspected at admission using transvaginal ultrasound (TVUS), performed by two expert sonographers from the Endometriosis Unit, following the IDEA and MUSA consensus. 8 , 9 However, the diagnosis was only established if these findings were confirmed by a second TVUS after clinical resolution of the acute episode and upon discharge. Data were retrospectively retrieved from patients' electronic medical records, including data on patient demographics and clinical characteristics (age, parity, ethnicity, pre‐existing medical conditions, previous abdominal surgery, and use of contraception at admission), as well as symptoms, blood analysis, and culture results. The TVUS imaging findings were collected at admission and at the one‐month follow‐up visit. For more accuracy, adnexal masses were classified into tubo‐ovarian abscess, pyosalpinx, and suspected infected ovarian endometrioma (Figure  1 ). Importantly, within the endometriosis group, we specifically distinguished between different subtypes of endometriosis, including adenomyosis, endometriomas, and DE (Figure  2 ). Adnexal ultrasound findings in patients with pelvic inflammatory disease. TVUS images showing characteristic adnexal findings: pyosalpinx (A), suspected infected endometrioma (B), tubo‐ovarian abscess (C). Transvaginal ultrasound appearance of endometriosis subtypes. TVUS images of the different endometriosis subtypes: Deep endometriosis (A), endometrioma (B), adenomyosis (C). We analyzed the following primary clinical outcomes: need for urgent surgical intervention during hospitalization, length of hospital stay, treatment failure, and PID recurrence rate. Empirical antibiotic treatment for admitted patients consisted of an intravenous third‐generation cephalosporin, intravenous metronidazole, and oral doxycycline in accordance with internal hospital guidelines. Treatment failure was defined as lack of improvement or clinical deterioration within 48–72 h post‐admission despite appropriate therapy. In such cases, antibiotics were escalated to broader‐spectrum agents (intravenous piperacillin‐tazobactam) or tailored based on culture results when available. Surgical intervention was considered in patients with treatment failure despite antibiotic therapy, clinical deterioration, persistent or enlarging adnexal collections, suspected abscess rupture, or when drainage or surgical management of the infectious focus was considered necessary by the attending surgical team. Surgeries were performed laparoscopically based on the findings: an abscess was drained and a tubo‐ovarian abscess was resected if necessary. Recurrence was defined as a new episode of PID requiring hospital readmission at least 30 days after discharge. This time threshold was established to differentiate recurrence from persistence of the initial infection. Finally, a subgroup analysis was performed within the endometriosis cohort to evaluate whether the presence of deep endometriosis, diagnosed by TVUS, was associated with worse clinical outcomes. Data analysis was conducted using descriptive statistics to summarize baseline characteristics and data retrieved at admission. Continuous variables are presented as mean and standard deviation or as median and interquartile range, as appropriate. Comparisons between patients with and without endometriosis were performed using the Student t ‐test for continuous variables and the chi‐square test for categorical variables. For the subgroup analysis comparing patients with and without deep endometriosis, the Mann‐Whitney U test and Fisher's exact test were used due to smaller sample size and distribution characteristics. To assess the independent effect of endometriosis on clinical outcomes, multivariate logistic and linear regression models were performed, adjusting for age, prior abdominal surgery, and hormonal treatment. Statistical significance was set at a p value of < 0.05. All statistical analyses were performed using Stata version 18.0 (StataCorp LP, College Station, TX, USA).