C-reactive protein and the menstrual cycle in females with sickle cell disease Open Access

Sickle cell disease (SCD) is a complex disorder that exhibits sex disparities in the frequency and severity of vaso-occlusive episodes (VOEs). Females with SCD experience more frequent VOEs with more severe pain, leading to increased hospitalizations, particularly during reproductive years, than male patients.^1,2 In addition, approximately half of all female patients with SCD report a temporal association between VOEs and their menstrual cycle, with VOEs clustering in the perimenstrual period.^3-5 This clinical pattern suggests that VOEs may be mediated by sex hormone cyclicity. In fact, limited studies demonstrate that certain hormonal contraceptives that suppress endogenous sex hormone fluctuation also have a therapeutic effect on SCD pain,^6,7 further supporting a hormonal component to vaso-occlusive pain, although the mechanism is poorly understood.

SCD is characterized by chronic inflammation stemming from hemolysis, endothelial dysfunction, and vaso-occlusion, with exacerbations during VOEs.^8,9 C-reactive protein (CRP) is a key inflammatory marker that is elevated at baseline in SCD^10,11 and rises acutely during VOEs.¹² In comparison with a panel of hematologic and inflammatory markers, including white blood cell count, hemoglobin, lactate dehydrogenase, vascular cell adhesion molecule-1, and P-selectin, CRP emerged as the most significant biomarker correlate of VOEs requiring hospitalization.¹³ Furthermore, this robust marker of inflammation in SCD has been shown to vary across the menstrual cycle in healthy females, peaking in the follicular phase.^14-17 

Whether CRP varies by sex or fluctuates with the menstrual cycle among female patients with SCD has not been previously investigated. Therefore, this study aims to explore patterns of CRP and other inflammatory markers among individuals with SCD.

We analyzed stored plasma samples with an associated diagnostic code for SCD from the Penn Medicine BioBank repository. Using the electronic medical record, we confirmed SCD diagnosis, genotype, and reproductive age (defined as 18-50 years old) at time of sample collection. We excluded participants who were pregnant, hospitalized with a VOE, or being treated at an emergency department or infusion center at time of sample collection. Demographics, medications, and laboratory information (including complete blood count and complete metabolic panel) from the time of sample collection were also abstracted from the electronic medical record.

We measured high sensitivity CRP in all samples and female sex hormones, including estradiol, progesterone, and luteinizing hormone, in samples from female patients. All laboratory work was performed using a Roche Cobas e411 Analyzer. Commercial enzyme-linked immunosorbent assays were used to quantify plasma levels of biomarkers of thrombosis (eg, thrombin-antithrombin complexes and von Willebrand factor; Abcam), neutrophil extracellular trap release (citrullinated histone 3; Cayman), and stress (cortisol; MP Biomedicals).

Given the absence of participant menstrual cycle information (eg, last menstrual period and menstrual cycle length) in the electronic medical record, we predefined sex hormone cutoffs to estimate menstrual cycle phase of female participants at time of sample collection. In keeping with typical clinical laboratory definitions and our internal research in predicting menstrual cycle phase from cross-sectional hormone values, we used a progesterone level of 1.75 ng/mL to define occurrence of ovulation and thus, the cutoff between follicular and luteal phases of the menstrual cycle.

We compared CRP, clinical laboratory markers, and other biomarker levels by patient sex, SCD genotype, and hydroxyurea use. Among female participant samples, we made the same comparisons between samples from the follicular and luteal phases of the menstrual cycle.

Statistical testing for comparison of medians between groups was performed using the Mann-Whitney and Kruskal-Wallis tests as appropriate.

This study was approved by the institutional review board of the University of Pennsylvania (protocol 851222) in accordance with the Declaration of Helsinki.

A total of 31 plasma samples from individuals with confirmed SCD met our inclusion criteria and were analyzed. Mean CRP concentration was 4.45 mg/L (standard deviation, 3.84 mg/mL). No significant differences were observed in median CRP levels by SS (n = 15) and SC (n = 13) genotype (4.59 mg/L [2.6-7.6] vs 5.21 mg/L [3.0-7.6]; P = .68); 2 patients had the S-β thalassemia genotype and 1 had an unknown genotype and were therefore, not included in this comparison. No difference was seen by treatment (n = 6) or lack of treatment (n = 25) with hydroxyurea (3.51 mg/L [1.6-7.5] vs 4.59 mg/L [2.4-7.2]; P = .54).

Participants categorized in the follicular phase had median estradiol 64.4 pg/mL, progesterone 0.42 ng/mL, and luteinizing hormone 11.4 IU/L; those in the luteal phase had estradiol 106.9 pg/mL, progesterone 7.34 ng/mL, and luteinizing hormone 4.48 IU/L. Median CRP levels were compared by female (n = 13) and male (n = 18) sex, with no difference being observed (3.88 mg/L [1.8-9.0] vs 4.45 mg/L [2.6-5.9]; P = .89). However, CRP did vary by menstrual cycle phase: female participants in the follicular phase had higher median CRP levels than those in the luteal phase (8.80 mg/L [2.7-10.5] vs 0.82 mg/L [0.6-2.1]; P = .03; Figure 1). Although median CRP did not differ across the sexes overall, there was a trend toward higher CRP among females in the follicular phase than in males (8.80 mg/L [2.7-10.5] vs 4.45 mg/L [2.6-5.9]; P = .59), and lower CRP among females in the luteal phase than in males (0.82 mg/L [0.6-2.1] vs 4.45 mg/L [2.6-5.9]; P = .22).

Neutrophil and platelet counts, aspartate aminotransferase, cortisol, and thrombin-antithrombin complexes also exhibited a trend toward elevation among females in the follicular phase compared with females in the luteal phase and to males. The median platelet counts in females in the follicular phase were significantly elevated compared to platelet counts in male patients (383 × 10³/μL [319 × 10³/μL to 463 × 10³/μL] vs 219 × 10³/μL [155 × 10³/μL to 269 × 10³/μL]; P < .001; Figure 2).

In this study, we observed a significant elevation in CRP in female patients with SCD in the follicular phase compared with those in the luteal phase of the menstrual cycle, suggesting a cyclical pattern of inflammation that peaks in the follicular phase. This trend mirrors CRP fluctuations across the menstrual cycle observed in the general population. However, the magnitude of CRP elevations in the follicular phase appears to be amplified in SCD, with the median plasma concentrations of 8.80 mg/L as observed in this study vs 0.74 mg/L in healthy females without SCD.¹⁴ Similarly, mean CRP levels in this SCD cohort were consistent with those reported in the literature^10,11 and higher than in the general population (eg, 1.31 mg/L in 1 study).¹⁰ Although previous studies have also shown that CRP levels are higher in individuals with SCD,^10,11 this is the first study, to our knowledge, to demonstrate that CRP levels in reproductive-aged females with SCD are significantly elevated during the follicular phase. CRP fluctuation across the menstrual cycle may have clinical implications in SCD, as VOEs exhibit a similar temporal pattern. An increased baseline CRP has been associated with increased frequency of VOE hospitalizations in children with SCD,¹³ demonstrating the potential clinical consequences of these biomarker fluctuations.

This study is limited by its small sample size and cross-sectional design. Each participant contributed only 1 study sample, thus biomarker differences attributed to menstrual cycle phase may represent other interindividual differences. We also lacked clinical menstrual cycle data among female patients; however, our use of sex hormone laboratory measurements still allowed for rigorous menstrual cycle phase definitions, while acknowledging the assumption of regular menstruation in our patients. We also acknowledge that females with SCD have increased prevalence of diminished ovarian reserve (DOR), which is associated with low sex hormone levels; follicle stimulating hormone levels would distinguish between DOR and follicular phase of the menstrual cycle, but these were not measured. Nevertheless, individuals with DOR often have very low estradiol (<20 pg/mL), which was not observed for any participant. In addition, the use of certain forms of hormonal contraception suppress endogenous sex hormone levels, which might have resulted in follicular phase misclassification according to our definitions. However, only 1 participant was documented as using hormonal contraception at the time of study collection, a progestin-containing intrauterine device, a method that does not completely suppress endogenous sex hormone cycling; this individual had an elevated progesterone level indicative of recent ovulation and was thus confidently classified in the luteal phase and kept in the study cohort. We also lacked information regarding the timing of blood draws, making it difficult to interpret differences in cortisol, which varies diurnally. Finally, because VOEs evolve through different stages and the prodromal phase is not typically associated with pain,¹⁸ we were unable to ensure that included patients, although not having symptoms of an acute VOE, were at their baseline inflammatory levels.

VOEs are a crucial target for intervention among individuals with SCD, yet perimenstrual VOE pathophysiology remains poorly understood. These results suggest a cyclic pattern of inflammation across the menstrual cycle in females with SCD that may contribute to perimenstrual VOEs. These findings warrant prospective validation, exploration of the menstrual patterns of other markers associated with SCD pathophysiology and correlation with clinical symptomatology. Further study of perimenstrual VOEs may support therapeutic strategies, including hormonal contraceptives,^6,7 outside of the typical armamentarium of SCD treatments.

Contribution: J.W., A.H.R., and K.G. were involved in project creation, data collection, data analysis, and manuscript writing; and V.B. was involved in data collection and data analysis.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Jessica Wu, Department of Obstetrics and Gynecology, Hospital of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104; email: jessica.wu@pennmedicine.upenn.edu.