AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 136:138–146 (2008) C-Reactive Protein Across the Menstrual Cycle Katherine Wander,1* Eleanor Brindle,2 and Kathleen A. O’Connor1,2 1 2 Department of Anthropology, University of Washington, Seattle, WA 98195 Center for Studies in Demography and Ecology, University of Washington, Seattle, WA 98195 KEY WORDS inﬂammation; estrogen; progesterone; biomarker ABSTRACT C-reactive protein (CRP) is a widely used, sensitive biomarker of inﬂammation. Studies conducted among users of exogenous hormones suggest that estrogen increases CRP, whereas progesterone decreases CRP. Examinations of CRP in normally cycling women suggest the opposite: CRP is negatively associated with endogenous estrogen and positively associated with endogenous progesterone. This work evaluates the association between menstrual cyclerelated hormone changes and events (menstruation and ovulation) and CRP. Eight female subjects gave urine and blood samples from twelve days across the menstrual cycle, for a total of eleven cycles. Blood samples were assayed for CRP; urine samples for b-follicle stimulating hormone (bFSH), pregnanediol 3-glucuronide (PDG), and estrone glucuronide (E1G). Ovulation day was estimated using hormone levels. Presence or absence of menses was reported by subjects. Analyses were conducted with random-effects linear regression. All cycles were ovulatory; day of ovulation was identiﬁed for nine cycles. A ten-fold increase in progesterone was associated with a 23% increase in CRP (P 5 0.01), a ten-fold increase in estrogen was associated with a 29% decrease in CRP (P 5 0.05), and menses was associated with a 17% increase in CRP (P 5 0.18); no association between ovulation or FSH and CRP was found. Hormone changes across the menstrual cycle should be controlled for in future studies of inﬂammation in reproductive-age women. Am J Phys Anthropol 136: 138–146, 2008. V 2008 Wiley-Liss, Inc. C-reactive protein (CRP) and other indicators of inﬂammation are widely used markers of health because they are both indicative of current infectious disease status and predictive of many chronic diseases. Inﬂammation is associated with a variety of chronic degenerative diseases, notably cardiovascular disease (CVD) (Danesh et al., 2000; Pearson et al., 2003) and adult-onset diabetes (Pradhan et al., 2001). Evidence is strong that the relationship between inﬂammation and chronic disease is not secondary to some other causal relationship, but that inﬂammation directly contributes to the damage responsible for these diseases (Ross, 1999; Libby et al., 2002; Dandona et al., 2004); however, the possibility remains that elevations in CRP result from, rather than contribute to, chronic disease (see, for example, Woodward et al., 2003). C-reactive protein is part of the acute phase immune response. Upon injury or infection, CRP levels can increase up to 10,000-fold (Pepys and Hirschfeld, 2003). This increase is rapid, but not immediate: circulating concentrations reach 5 mg/L by about 6 h and peak by about 48 h; the half-life of CRP in the blood is around 19 h (Vigushin et al., 1993). Because production of CRP responds predictably to insult, quickly returns to normal following insult, and clearance from the blood happens at a constant rate, CRP levels provide a reliable index of ongoing inﬂammation (Vigushin et al., 1993; Pepys and Hirschfeld, 2003). Dramatic elevations in CRP concentration indicate ongoing infection; absent infection, CRP reﬂects the degree of background inﬂammation characteristic of an individual (Pepys and Hirschfeld, 2003). The utility of CRP for anthropological, epidemiological, and nutritional research is quite broad. Using CRP, myriad studies have demonstrated a relationship between inﬂammation and degenerative disease in Western popu- lations (i.e., Danesh et al., 2000; Pradhan et al., 2001). This relationship has recently been investigated among a traditionally nonagricultural population as well: among the Yakut, Sorensen et al. (2006) demonstrate low CRP values (relative to Western populations), despite the presence of substantial CVD mortality, suggesting that subsistence activities may modify the relationship between inﬂammation and CVD. Beyond investigations of degenerative disease, CRP is used as a marker of acute infection. Identifying acute infection is crucial to accurate characterization of health and nutritional status (i.e., Beall et al., 2002; Shell-Duncan and McDade, 2004), as the acute phase response substantially alters expression of many biomarkers. CRP and other acute phase proteins have also been used to assess environmental and infectious disease stresses among non-Western children (Panter-Brick et al., 2001; McDade et al., 2005). Here, we explore the effect of the menstrual cycle on CRP, to inform future use of this biomarker and to shed light on interactions between human reproduction and immune function. C 2008 V WILEY-LISS, INC. C Grant sponsors: National Science Foundation Graduate Research Fellowship, National Institute of Child Health and Human Development; Grant number: R24D042828. *Correspondence to: Katherine Wander, Department of Anthropology, Box 353100, University of Washington, Seattle, Washington 98195, USA. E-mail: email@example.com Received 16 July 2007; accepted 29 November 2007 DOI 10.1002/ajpa.20785 Published online 6 February 2008 in Wiley InterScience (www.interscience.wiley.com). CRP ACROSS THE MENSTRUAL CYCLE CORRELATES OF C-REACTIVE PROTEIN Within individuals, CRP measures are moderately stable over time: when classiﬁed by CRP (i.e., the American Heart Association’s suggestions for classifying CVD risk, Pearson et al., 2003), about two thirds fall into the same category when reclassiﬁed across the course of 1 year (see Macy et al., 1997; Kluft and de Maat, 2001; Ockene et al., 2001). In addition to injury and infection, strenuous exercise elevates individual CRP (Brull et al., 2003); moderate alcohol consumption (Sierksma et al., 2002) and weight loss (Tchernof et al., 2002) decrease CRP. Statin therapy (Albert et al., 2001) and possibly aspirin (Ikonomidis et al., 1999) reduce CRP; few other pharmaceuticals are known to alter CRP levels. There is no systematic diurnal (Meier-Ewert et al., 2001) or seasonal (Fröhlich et al., 2002) variation in CRP. Among individuals not experiencing infection, CRP concentrations vary consistently with many attributes. In adults, CRP is positively associated with age (Chenillot et al., 2000; Ford et al., 2003) and negatively associated with birth weight (Sattar et al., 2004). Adult females have higher CRP than males (Wener et al., 2000; Ford et al., 2003). Smokers have higher CRP than nonsmokers (Chenillot et al., 2000). CRP is negatively associated with measures of physical ﬁtness (LaMonte et al., 2002), and positively associated with BMI and adiposity (Visser et al., 1999; LaMonte et al., 2002; Tchernof et al., 2002). Some component of variation in circulating levels of CRP is heritable: estimates of heritability between 30% (Austin et al., 2004) and 50% (MacGregor et al., 2004) have been offered. Oral contraceptive (OC) use (Dreon et al., 2002; Kluft et al., 2002) and orally administered hormone replacement therapy (HRT) (Walsh et al., 2000; Kluft et al., 2002) consistently elevate CRP, whether the preparation includes progesterone and estrogen or estrogen alone. Estrogen administration is also associated with increases in CRP among men being treated for prostate cancer (Kovacs et al., 2005). Such studies repeatedly demonstrate that exogenous estrogen can increase CRP. However, dose and delivery method appear to mediate exogenous estrogen’s effect on CRP: Prestwood et al. (2004) demonstrate signiﬁcant increases in CRP among women taking a higher estrogen HRT preparation and signiﬁcant decreases among women taking a lower estrogen preparation; transdermally administered HRT has no demonstrable effect on CRP (Decensi et al., 2002; Ropponen et al., 2005). Comparisons of different HRT preparations suggest that exogenous progesterone decreases CRP (Cushman et al., 1999; Skouby et al., 2002; but see Ridker et al., 1999). A pro-inﬂammatory effect of exogenous progesterone administration has been reported in men (Zitzmann et al., 2005). The use of progesterones of different types complicates these studies, as progestational agents differ in receptor afﬁnity and glucocorticoid effects; the ‘‘pure progestational’’ Nomegestrol acetate (which interacts only with progesterone receptors and has no glucocorticoid effects) was found to be associated with decreases in CRP (Gol et al., 2006). The results of the few existing studies examining the effects of endogenous hormones on CRP appear to be inconsistent with the above trends. Jilma et al. (1997) examined measures from 18 women taken in the follicular phase, at mid-cycle, and in the luteal phase of one menstrual cycle, and found a signiﬁcant (univariate) 139 association between individuals’ relative increase in progesterone and CRP from the follicular phase to mid-cycle (r 5 0.60, P 5 0.01), as well as from the follicular phase to the luteal phase (r 5 0.71, P 5 0.001). Blum et al. (2005) examined 15 measures from 15 women across one menstrual cycle. In random effects maximum likelihood regression, increasing estrogen was associated with decreasing CRP (b 5 20.23, P < 0.001; univariate) such that a ten-fold increase in estrogen was associated with a 41% decrease in CRP. The effect of progesterone on CRP in this study was not signiﬁcant (P 5 0.15). Although subjects were sampled daily around ovulation, a potential effect of ovulation per se on CRP concentration was not discussed. In this study, CRP concentrations were highest during menstruation; independent effects of estrogen and menses were not tested. Thus, the effects of endogenous and exogenous hormones on CRP appear to differ: oral exogenous estrogen at most doses is associated with elevations in CRP (Walsh et al., 2000; Dreon et al., 2002; Kluft et al., 2002; Kovacs et al., 2005), endogenous estrogen seems to be associated with decreases in CRP (Blum et al., 2005); exogenous progesterone may be associated with decreases in CRP (Cushman et al., 1999; Skouby et al., 2002; Gol et al., 2006), endogenous progesterone seems to be associated with increases in CRP (Jilma et al., 1997). Neither study examining the effects of endogenous hormones reports any evaluation of confounding between the effects of estrogen and progesterone, despite the fact that changes in these hormones regularly co-occur across the menstrual cycle. Further, these studies do not evaluate potential independent effects of menses or ovulation, despite good reason to expect these events to substantially affect inﬂammation. Ovulation and menstruation have both been described as ‘‘inﬂammatory’’ events (Espey, 1980; Kelly, 1994). Inﬂammatory mediators and leukocytes play important roles in folliculogenesis and follicular rupture (Bukulmez and Arici, 2000). Similarly, prostaglandins and leukocytes play a direct role in accomplishing menstruation (Critchley et al., 2001). As such, an effect of these menstrual cycle events on circulating concentrations of CRP bears exploration. Further, because elevations in follicle stimulating hormone (FSH) are typically associated with both menstruation (Miro and Aspinall, 2005) and ovulation (Schwartz, 1974), it is possible that changes in this hormone may explain any association between those cycle events and CRP. Thus, this pilot study was undertaken to verify the unexpected effects of endogenous reproductive steroids on CRP, to evaluate potential effects of menses, ovulation, and FSH on CRP, and to evaluate potential confounding and interaction in these relationships. METHODS Participants and specimen collection Participants were eight non-obese normally cycling females between 21 and 47 years of age without any reported history of diabetes or cardiovascular disease; none used exogenous hormones, statins, or more than occasional anti-inﬂammatory medication. Paired urine and capillary dried blood spot (DBS) collections were made from three volunteers for two months (24 collections per subject), and from ﬁve volunteers for one month (12 collections per subject). Collections were made every Monday, Wednesday, and Friday during the collecAmerican Journal of Physical Anthropology 140 K. WANDER ET AL. tion period. Procedures were approved by the Institutional Review Board of the University of Washington. Although venous blood is more commonly used in examinations of CRP as a biomarker, DBS were used in this study for two reasons: this collection method is minimally invasive and was believed to be more conducive to collection of serial samples; further, DBS are more likely to be used in examinations of CRP in populations of interest to anthropologists (McDade et al., 2004). Blood spots were collected by ﬁnger stick with a sterile lancet. Spots of blood were allowed to fall freely from the subject’s ﬁnger onto ﬁlter paper (Schleicher and Schuell #903 Specimen Collection Paper). Filter papers were allowed to dry at room temperature in open air for one to three days, and were then frozen with desiccant until assay. Within 1 h of DBS collection, subjects made urine collections; urine was collected and refrigerated in prelabeled plastic vials. Most subjects made urine collections in the laboratory; other subjects collected and refrigerated urine samples in their homes for later transport to the laboratory. In the laboratory, urine-speciﬁc gravity measures were taken and a 600 lL aliquot was made and refrigerated until assay; the remaining sample was frozen. All initial assays were completed within two weeks of the completion of subjects’ sample collection periods; later reanalysis of some samples was conducted using stored sample. To address the possibility that DBS ﬁnger stick collection procedures were invasive enough to evoke some measure of inﬂammatory response, subjects’ start days were assigned irrespective of cycle day, so that no collection day systematically co-occurred with any cycle day. Further, collections were not made over the weekend. If collection procedures induced any increase in CRP, the one-day interval preceding Wednesday and Friday collections would allow for more limited clearance than the two-day interval preceding Monday collections, resulting in regularly lower CRP values on Mondays. Thus, a test for systematically lower CRP on Mondays could be employed to evaluate a potential effect of ﬁnger-stick procedures on CRP. Assays Dried blood spots were assayed for CRP with a highsensitivity sandwich enzyme immunoassay using polyclonal antibodies, described in detail elsewhere (McDade et al., 2004). Microtiter plates were coated with antiCRP antibody (DAKO, cat. no. A0073, recently discontinued). Calibrators (Fitzgerald Industries International, Inc., cat. no. 30-AC10; combined with washed erythrocytes and preserved on ﬁlter paper), in house controls, and specimens were eluted in buffer from ﬁlter paper. Horseradish peroxidase-labeled detection antibody (DAKO, cat. no. P0227, recently discontinued) and substrate solution were used to develop a color signal proportional to CRP concentration. DBS measures of CRP using this assay correlate highly with serum measures of CRP (r 5 0.96 for 84 paired DBS and serum specimens; McDade et al., 2004). For the eight plates run for this project, the inter-assay coefﬁcient of variation (CV) for in-house controls was 2.9% at high (3.0 mg/L) and 15.8% at low (0.11 mg/L) concentration. Urine samples were assayed for estrone glucuronide (E1G) and pregnanediol 3-glucuronide (PDG), the major urinary metabolites of estradiol (E2) and progesterone American Journal of Physical Anthropology (P4), respectively. Competitive enzyme immunoassays (EIAs) used to quantify urinary levels of E1G and PDG are described in detail elsewhere (O’Connor et al., 2003, 2004). Plates were coated monoclonal capture antibodies (Quidel Corporation, clone 330 for PDG; Weizmann Institute, clone 3F11 for E1G). Urine specimens, calibrators (Sigma, cat. no. P3635 for PDG; Sigma, cat. no. E1752 for E1G), in-house controls, and PDG or E1G conjugated to horseradish peroxidase were added to the wells. Substrate was added to develop a color signal inversely proportional to steroid conjugate concentration. Urine samples were assayed for FSH using a sandwich immunoenzymometric assay speciﬁc to total (free and intact) bFSH (Brindle et al., 2006). Microtiter plates were coated with a monoclonal capture antibody speciﬁc to bFSH (Scantibodies Laboratory, Inc., clone FS2.4A10.G10). Urine samples, in-house controls, and calibrators (bFSH, A. F. Parlow, AFP2911A, NIDDK NHPP) were added to the wells, followed by polyclonal anti-bFSH detection antibody (A. F. Parlow, NIDDKanti-hBetaFSH-1, NHPP). Addition of horseradish peroxidase-labeled antibody and substrate solution yielded a color signal proportional to FSH concentration. Urinary PDG, E1G, and bFSH measured using these assays parallel serum progesterone, estradiol, and intact FSH, respectively, across the menstrual cycle: Pearson correlations between paired urine and serum specimens from 30 averaged cycles (n 5 34 cycle days) were 0.98 for PDG-progesterone (O’Connor et al., 2003), 0.94 for E1G-estradiol (O’Connor et al., 2004), and 0.86 for bFSH-serum intact FSH (Brindle et al., 2006). For quality control, assay runs were repeated if the estimated value of both high and low controls fell outside the controls’ established 95% conﬁdence interval; control CI’s are calculated from 47 and 48 plates for E1G, 38 and 39 plates for PDG, and 3,352 plates for FSH. All specimens, calibrators and controls were run in duplicate; concentrations were estimated from optical density using a four-parameter logistic model in BIOLINX 2.0 Software (Dynatech Laboratories, Chantilly, VA). Urinary hormone values were adjusted for variation in hydration status by speciﬁc gravity (Miller et al., 2004); E1G results were statistically corrected for slight assay nonparallelism (O’Connor et al., 2004). Identiﬁcation of ovulation Cycles were identiﬁed as ovulatory or anovulatory and day of ovulation was identiﬁed using methods developed for intermittent sampling (O’Connor et al., 2006), which rely on changes in urinary measures of steroid hormones. A cycle was identiﬁed as ovulatory if it evinced a relative rise in progesterone (as reﬂected by the running average PDG for each collection) that was sustained for at least two consecutive collections. Ovulation day was identiﬁed as the second day of a ﬁve-day dramatic decline in the ratio of estrogen to progesterone. When adapted for intermittent sampling, this method (modiﬁed Baird) identiﬁed ovulation day to within two days of the true ovulation day for 93% of cycles. When a noncollection day was identiﬁed as the day of ovulation, the following collection was deﬁned as ovulation day for these analyses. Sampling in this study provided slightly fewer samples than that used to develop these methods, so the decision rule was relaxed regarding the change in the ratio of steroid hormones necessary to identify ovulation. The protocol calls for a 141 CRP ACROSS THE MENSTRUAL CYCLE Fig. 1. Variation in hormones and CRP across the menstrual cycle: The nine cycles for which ovulation day could be estimated were centered on the estimated ovulation collection, and the mean value was calculated for the ovulation collection, the preceding six collections, and the ﬁve following collections. Between six and nine values are included in the mean for each collection. Cycles were not centered on menstruation; however, menses was reported by most participants at the sixth or ﬁfth collection preceding estimated ovulation. decrease of 60% across a ﬁve-day window. Since cycles clearly identiﬁed as ovulatory did not consistently demonstrate this magnitude decrease, but did demonstrate a 50% decrease, a 50% decrease was accepted in identifying ovulation day in these analyses. Statistical analysis Data were analyzed using StataSE 9.0 (Statacorp, College Station, TX). Statistical signiﬁcance was deﬁned as P 0.05; suggestive results were deﬁned as 0.05 < P 0.20; confounding was deﬁned as a change >10% in the coefﬁcient of the predictor of interest with the inclusion of a potential confounder. Relationships between CRP and other variables were analyzed using random-effects linear regression, wherein a random intercept term was estimated for each participant to account for differences among individuals. To assess the effect of infection-related changes in CRP, the magnitude of which could obscure meaningful changes in background CRP, and which could introduce potential confounding, analyses were carried out both with and without data points for which CRP was elevated. Choice of such cut-off varies depending on the ﬂuid analyzed and the type and severity of infection identiﬁed. Here, we employ the cut-off CRP > 5 mg/L (238 mmol/L). Because previous work (Prestwood et al., 2004) suggests opposite effects of low and high dose estrogen on CRP, we chose to test the hypothesis that the effects of estrogen on CRP may change at high concentrations. To do this, analyses were carried out both with and without points above the 90th percentile E1G. RESULTS Data from eleven menstrual cycles of eight female subjects were used to assess menstrual cycle-related changes in CRP. Values for E1G and PDG were available for a total of 126 data points; values for FSH were available for 124 data points. The presence or absence of menses was reported by participants for 126 data points; menses was reported as present for 37 data points. All cycles were ovulatory; day of ovulation was identiﬁable for nine cycles. Variation in hormones and CRP across these cycles (centered on ovulation) is presented in Figure 1. CRP, PDG, E1G, and FSH showed skewed distributions and were therefore transformed using a natural logarithm; within individuals, transformed CRP values appear normally distributed. No subjects reported using more than occasional anti-inﬂammatory medication. The majority of anti-inﬂammatory medications reported were ibuprofen and acetaminophen. On one occasion, one subject reported using aspirin; the associated collection was dropped from all analyses. Collection procedures did not artiﬁcially elevate CRP: Monday collections (including all collections preceded by two or more noncollection days) were tested for systematically lower CRP concentrations within individuals; no effect was found (b 5 0.085, P 5 0.26). Confounding was apparent between all predictors of interest for which there was a suggestive univariate effect (Table 1); only one model was ultimately used for American Journal of Physical Anthropology 142 K. WANDER ET AL. TABLE 1. Evaluation of confounding Univariate associations with C-reactive protein, ln(CRP) [Coefﬁcient (P-value)] Progesterone, ln(PDG) 0.030 (0.33) Estrogen, ln(E1G) 20.093 (0.16) Menses 0.179 (0.10) Adjusted for potential confounders (bivariate) ln(PDG) ln(E1G) Menses ln(FSH) Ovulation – 0.088 (0.01) 0.052 (0.07) 0.027 (0.40) 0.032 (0.29) 20.191 (0.01) – 20.045 (0.49) 20.096 (0.15) 20.080 (0.24) 0.214 (0.06) 0.154 (0.18) – 0.169 (0.14) 0.169 (0.13) Results of random-effects linear regression show substantial confounding between these three predictors of interest. CRP, Creactive protein; PDG, pregnanediol 3 glucuronide, reﬂects serum progesterone levels; E1G, estrone glucuronide, reﬂects serum estradiol levels. The top row lists the coefﬁcients and P-values for univariate associations between a variable and CRP, the lower rows show how those coefﬁcients and P-values change with the inclusion of each potential confounder (listed to the left). Thus, for example, the coefﬁcient for estrogen increases in magnitude, from 20.093 to 20.191, and signiﬁcance, from P 5 0.16 to P 5 0.01, with the inclusion of progesterone in the model. all predictors (Table 2). All possible two-way interaction terms from this model were tested individually. None were found to be signiﬁcant; however, the interaction term coefﬁcients for ln(PDG)*ln(E1G) (P 5 0.06) and ln(E1G)*menses (P 5 0.18) were suggestive. Thus, future studies with larger sample sizes should assess interaction between the effects of progesterone and estrogen and the effects of estrogen and menses on CRP. According to the model presented in Table 2, controlling for E1G and menses, a ten-fold increase (typical for steroids across the cycles we observed) in PDG is associated with a 23% increase in CRP (P 5 0.01). Controlling for PDG and menses, a ten-fold increase in E1G is associated with a 29% decrease in CRP (P 5 0.05). Controlling for these hormones, a suggestive increase of 17% in CRP (P 5 0.18) is associated with the presence of menses. R-squared values for this model are low, indicating that a small proportion of the variation among normally cycling women in CRP is explained by cycle-related variables; however, a signiﬁcant proportion of the variation in CRP within a normally cycling woman is explained by these variables. The effects of steroid hormones on CRP were assessed within each cycle phase through two separate regression models; menses was not included in either of these models (Table 3). Within the follicular phase, the effect of estrogen is of greater magnitude and signiﬁcance than across the entire cycle, despite a reduction in sample size, while the effect of progesterone is not apparent. Within the luteal phase, the effect of progesterone is apparent and signiﬁcant, despite the reduced sample size; the effect of estrogen is of lesser magnitude and signiﬁcance than observed in either the follicular phase or the entire cycle. No association between FSH and CRP, nor any confounding between FSH and other predictors, was found. Thus, FSH was disregarded in further analyses. The expected increase in FSH with menses (b 5 0.311, P 5 0.01) and ovulation (b 5 0.679, P 5 0.00) was evident. No association between ovulation and CRP (b 5 20.055, P 5 0.60 when included in the model in Table 2), nor any confounding between ovulation and the effect American Journal of Physical Anthropology of steroids on CRP, was found. Thus, ovulation was disregarded in further analyses. Four CRP values exceeded 5 mg/L (238 mmol/L); their exclusion from the ﬁnal regression model does not substantially alter the magnitude of any coefﬁcients; the coefﬁcient for E1G, however, loses signiﬁcance without these data points (Table 4). When the highest E1G values are excluded from the analysis, the negative relationship between E1G and CRP becomes stronger, despite the reduction in sample size. The results of the regression model excluding points with E1G above the 90th percentile [E1G 101,675.0 pg/mL (47.63 mmol/L), 13 values] are shown in Table 5. Compared with the results in Table 2, coefﬁcients for PDG and menses decrease in magnitude and signiﬁcance, concomitant with a reduction in sample size; the E1G coefﬁcient increases in magnitude and signiﬁcance, such that a ten-fold increase in E1G is associated with a 43% decrease in CRP (P 5 0.02). DISCUSSION Within individual women, CRP concentrations are clearly associated with endogenous reproductive hormones: increases in estrogen are associated with decreases in CRP (P 5 0.05), while increases in progesterone are associated with increases in CRP (P 5 0.01). Independent of hormone changes, the presence of menses is associated with nonsigniﬁcant increases in CRP (P 5 0.18). Substantial confounding is apparent between these effects; this may explain the lower magnitude of the effects reported here, compared to previous studies (Jilma et al., 1997; Blum et al., 2005). The negative association between estrogen and CRP is stronger in the follicular phase, while the positive association between progesterone and CRP is limited to the luteal phase. Estrogen exhibits opposing effects on a number of physiological variables at low and high levels (Calabrese, 2001). Evidence suggests that this pattern holds for inﬂammation as well: increases in exogenous estrogen are negatively associated in vivo with circulating CRP (Prestwood et al., 2004) and in vitro with whole blood interleukin (IL)-1b and IL-6 production (Rogers and Eastell, 2001) at lower estrogen doses, and positively associated at higher estrogen doses. When we excluded the highest 10% of estrogen values (those that might be proinﬂammatory), the negative effect of estrogen was strengthened. This supports the hypothesis that estrogen has opposite effects on inﬂammation at high and low levels. Our ﬁnding of a negative relationship between CRP and endogenous estrogen at most levels is consistent with the results reported by Blum et al. (2005); these results are in contrast to the ﬁndings of studies focusing on exogenous estrogen (Dreon et al., 2002; Kluft et al., 2002). The opposite effects of exogenous and endogenous estrogen may result in part from changes in the effect of estrogen at high and low doses (described above). They may also be the result of ﬁrst-pass effects of oral estrogen on hepatic CRP production (Decensi et al., 2002). Our results indicate that increases in estrogen at levels found across most of the menstrual cycle are associated with decreases in CRP; and that changes in the relationship between estrogen and inﬂammation occur across normal physiological levels of estrogen. An antiinﬂammatory effect of increasing estrogen concentrations at most physiological levels is consistent with estrogen’s 143 CRP ACROSS THE MENSTRUAL CYCLE TABLE 2. Relationship between CRP and menstrual cycle variables Outcome variable: ln(CRP) Predictor variable Progesterone, ln(PDG) Estrogen, ln(E1G) Menses Coefﬁcient 0.090 20.148 0.156 P-value 0.01 0.05 0.18 95% CI 0.023, 0.157 20.297, 0.000 20.074, 0.387 Interpretation 100.090 5 1.23 1020.148 5 0.711 e0.156 5 1.17 R2 within 5 0.0754, R2 between 5 0.0005, R2 overall 5 0.0053. TABLE 3. Relationship between CRP and steroid hormones by cycle phase Outcome variable: ln(CRP) Predictor variable Follicular phase Progesterone, ln(PDG) Estrogen, ln(E1G) Luteal phase Progesterone, ln(PDG) Estrogen, ln(E1G) Coefﬁcient P-value Interpretation 0.021 20.285 0.84 0.01 100.021 5 1.05 1020.285 5 0.519 0.05 0.32 5 1.30 10 1020.106 5 0.783 0.114 20.106 Outcome variable: ln(CRP) Coefﬁcient 0.081 20.107 0.182 Outcome variable: ln(CRP) Predictor variable Progesterone, ln(PDG) Estrogen, ln(E1G) Menses Coefﬁcient 0.088 20.242 0.099 P-value 0.04 0.02 0.44 Interpretation 100.088 5 1.22 1020.242 5 0.573 e0.099 5 1.10 R2 within 5 0.0793, R2 between 5 0.0095, R2 overall 5 0.0154. 0.114 TABLE 4. Effect of excluding high CRP values Predictor variable Progesterone, ln(PDG) Estrogen, ln(E1G) Menses TABLE 5. Effect of excluding high estrogen values P-value Interpretation 0.01 100.081 5 1.21 0.12 1020.107 5 0.782 0.11 e0.182 5 1.20 R2 within 5 0.0867, R2 between 5 0.0368, R2 overall 5 0.0001. effect on inﬂammation in many tissue types: estrogen has a consistently negative effect on inﬂammatory cell migration and inﬂammatory marker production in diverse nonreproductive, nonimmune tissues (Girasole et al., 1992; Ashcroft et al., 1999; Miyamoto et al., 1999; Vegeto et al., 2001). It is possible that increases in inﬂammation with the highest levels of estrogen indicate that estrogen plays a role in accomplishing increases in inﬂammation observed across pregnancy (Watts et al., 1991; Sacks et al., 2004). Our ﬁnding of a positive effect of progesterone on CRP is consistent with the relationship reported by Jilma et al. (1997). It is in contrast to the negative effect suggested by studies of exogenous hormone administration; however, these studies are themselves quite inconsistent (for discussions, see Ropponen et al., 2005; Gol et al., 2006). Our ﬁnding is difﬁcult to reconcile with many observations of progesterone’s effects on the immune system: while progesterone seems to promote the chemotactic activity of neutrophils and increase production of some inﬂammatory mediators (IL-6 and leukemia inhibiting factor, LIF) by monocytes (Bouman et al., 2005), an overall anti-inﬂammatory effect is generally attributed to progesterone, which includes decreasing NK cell activity and macrophage tumor necrosis factor (TNF) a and nitric oxide synthase (NOS) production, as well as inhibiting T-cell development in the thymus and T-cell activity (for recent review, see Roberts et al., 2001). Our results suggest that increases in progesterone typical of normally cycling women have an overall inﬂammatory effect. Future work should focus on reconciling the observed effects of progesterone on inﬂammatory cells with its overall effect on systemic inﬂammation. We report that menstruation is associated with a suggestive increase in CRP; this is consistent with Blum et al. (2005) observation of highest CRP concentrations during menstruation. Because of its small sample size, our analysis lacks the power to reject alternative hypotheses at the 0.05 a-level for the observed magnitude of the effect of menstruation; thus we interpret this result as a suggestive pro-inﬂammatory effect of menstruation. This uncontrolled effect is clear in our data (see Fig. 1 and Table 1), and the suggestive effect persists after controlling for the confounding effects of hormones. Studies with larger samples should anticipate a proinﬂammatory effect of menses, consistent with the innate immune system’s role in accomplishing menstruation (Critchley et al., 2001). Despite the involvement of inﬂammatory cells and mediators in follicular rupture, ovulation does not appear to affect systemic inﬂammation. The possibility exists that our methods of identifying ovulation, sampling frequency, or power were inadequate to detect such an effect, but the small magnitude and high P-value of the coefﬁcient that results from ovulation’s inclusion in our model suggest that this is unlikely to be the case. CONCLUSIONS AND IMPLICATIONS Changes in estrogen and progesterone concentration are signiﬁcantly and independently associated with changes in CRP concentration, suggesting that progesterone increases CRP, estrogen decreases CRP, and menstruation may increase CRP. Overall, these changes are consistent with observed interactions between the reproductive and immune systems. The patterns we observed are not consistent with reported effects of exogenous hormones on CRP; our results agree with others examining endogenous reproductive steroids and CRP (Jilma et al., 1997; Blum et al., 2005), and our analysis expands on them, providing a single model of the independent associations between endogenous hormones and menstrual cycle events and CRP. C-reactive protein’s utility to human biology research is only beginning to be realized. In addition to nutritional anthropology and infectious disease research, investigations of human senescence and life history evolution will beneﬁt from using CRP to explore variation in background inﬂammation, its determinants, its effect on life expectancy, and its contribution to post-reproductive mortality. As investigations of inﬂammation across American Journal of Physical Anthropology 144 K. WANDER ET AL. the lifecourse are undertaken, it will be important to evaluate and control for potential confounding by reproductive variables. We suggest that studies using CRP as a biomarker among reproductive-age women account for cycle-related hormonal changes and menstruation in sampling and analysis; further, when study samples include individuals who may be undergoing any transition in their reproductive life course (adolescents, mothers returning to menses postpartum, or perimenopausal women), differences in hormone levels or frequency of menses may confound estimated effects of other variables on CRP. For example, Mexican-American children included in NHANES III have higher CRP than white children (controlling for age and BMI), especially among girls (3–17 years; Ford, 2003). This result may in part be attributable to earlier menarche among Mexican–American girls (Wu et al., 2002), which would result in the inclusion of more menses-associated collections for Mexican–American, compared to white, girls, elevating the estimated mean CRP among Mexican–American girls. Although it is unlikely that earlier menarche could entirely account for the observed pattern, we nonetheless suggest controlling for the presence of menses in future examinations of CRP among this age group. Similarly, variation in CRP across adult populations may be attributable in part to variation in steroid hormone levels. Forouhi et al. (2001) report signiﬁcantly higher CRP among South Asian women living in London, compared to white women of the same BMI category. Signiﬁcant differences between these populations in both progesterone and estrogen levels (O’Connor et al., 2003), which persist beyond migration (Núñez-de la Mora et al., 2007), may contribute to these differences in CRP. Thus, we suggest controlling for steroid hormones levels in future comparisons of CRP and inﬂammation across populations. ACKNOWLEDGMENTS The authors thank Jennifer Aranda, Masako Fujita, Steven Goodreau, Anita Rocha, Bettina Shell-Duncan, Jane Shofer, and anonymous reviewers for their help in the preparation of this manuscript. LITERATURE CITED Albert MA, Danielson E, Rifai N, Ridker PM. 2001. Effect of statin therapy on C-reactive protein levels: the Pravastatin Inﬂammation/CRP Evaluation (PRINCE): a randomized trial and cohort study. JAMA 286:64–70. Ashcroft GS, Greenwell-Wild T, Horan MA, Wahi SM, Ferguson MWJ. 1999. Topical estrogen accelerates cutaneous wound healing in aged humans associated with an altered inﬂammatory response. Am J Pathol 155:1137–1146. Austin MA, Zhang C, Humphries SE, Chandler WL, Talmud PJ, Edwards KL, Leonetti DL, Mcneely MJ, Fujimoto WY. 2004. Heritability of C-reactive protein and association with apolipoprotein E genotypes in Japanese Americans. Ann Hum Genet 68:179–188. Beall CM, Decker MJ, Brittenham GM, Kushner I, Gebremedhin A, Strohl KP. 2002. An Ethiopian pattern of human adaptation to high-altitude hypoxia. Proc Natl Acad Sci USA 99:17215–17218. Blum CA, Müller B, Huber P, Kraenzlin M, Schindler C, De Geyter C, Keller U, Puder JJ. 2005. Low-grade inﬂammation and estimates of insulin resistance during the menstrual cycle in lean and overweight women. J Clin Endocrinol Metab 90: 3230–3235. American Journal of Physical Anthropology Bouman A, Heineman MJ, Faas MM. 2005. Sex hormones and the immune response in humans. Hum Reprod Update 11: 411–425. Brindle E, Miller RC, Shofer JB, Klein NA, Soules MR, O’Connor KA. 2006. Urinary beta-luteinizing hormone and beta-follicle stimulating hormone immunoenzymometric assays for population research. Clin Biochem 39:1071–1079. Brull DJ, Serrano N, Zito F, Jones L, Montgomery HE, Rumley A, Sharma P, Lowe GDO, World MJ, Humphries SE, Hingorani AD. 2003. Human CRP gene polymorphism inﬂuences CRP levels. Arterioscler Thromb Vasc Biol 23:2063–2069. Bukulmez O, Arici A. 2000. Leukocytes in ovarian function. Hum Reprod Update 6:1–15. Calabrese EJ. 2001. Estrogen and related compounds: biphasic dose responses. Crit Rev Toxicol 31:503–515. Chenillot O, Henny J, Steinmetz J, Herbeth B, Wagner C, Siest G. 2000. High sensitivity C-reactive protein: biological variations and reference limits. Clin Chem Lab Med 38:1003–1011. Critchley HOD, Kelly RW, Brenner RM, Baird DT. 2001. The endocrinology of menstruation-a role for the immune system. Clin Endocrinol 55:701–710. Cushman M, Mailahn EN, Psaty BM, Kuller LH, Dobs AS, Tracy RP. 1999. Hormone replacement therapy, inﬂammation and hemostasis in elderly women. Arterioscler Thromb Vasc Biol 19:893–899. Dandona P, Aljada A, Bandyopadhyay A. 2004. Inﬂammation: the link between insulin resistance, obesity, and diabetes. Trends Immunol 25:4–7. Danesh J, Whincup P, Walker M, Lennon L, Thomson A, Appleby P, Gallimore JR, Pepys MB. 2000. Low grade inﬂammation and coronary heart disease: prospective study and updated meta-analyses. Br Med J 321:199–204. Decensi A, Omodei U, Robertson C, Bonanni B, Guerrieri-gonzaga A, Ramazzotto F, Johansson H, Mora S, Sandri MT, Cazzaniga M, Franchi M, Pecorelli S. 2002. Effect of transdermal estradiol and oral conjugated estrogen on C-reactive protein in retinoid-placebo trial in healthy women. Circulation 106:1224–1228. Dreon DM, Slavin JL, Phinney SD. 2002. Oral contraceptive use and increased plasma concentration of C-reactive protein. Life Sci 73:1245–1252. Espey LL. 1980. Ovulation as an inﬂammatory reaction-a hypothesis. Biol Reprod 22:73–106. Ford ES. 2003. C-reactive protein concentration and cardiovascular disease risk factors in children: ﬁndings from the National Health and Nutrition Examination Survey 1999– 2000. Circulation 108:1053–1058. Ford ES, Giles WH, Myers GL, Rifal N, Ridker PM, Mannino DM. 2003. C-reactive protein concentration distribution among US children and young adults: ﬁndings from the National Health and Nutrition Examination Survey, 1999–2000. Clin Chem 49:1353–1357. Forouhi NG, Sattar N, McKeigue PM. 2001. Relation of C-reactive protein to body fat distribution and features of the metabolic syndrome in Europeans and South Asians. Int J Obes 25:1327–1331. Fröhlich M, Sund M, Thorand B, Hutchinson WL, Pepys MB, Koenig W. 2002. Lack of seasonal variation in C-reactive protein. Clin Chem 48:575–577. Girasole G, Jilka RL, Passeri G, Boswell S, Boder G, Williams DC, Manolagas SC. 1992. 17b-Estradiol inhibits interleukin-6 production by bone marrow-derived stromal cells and osteoblasts in vitro: a potential mechanism for the antiosteoporotic effect of estrogens. J Clin Invest 89:883–891. Gol M, Akan P, Dogan E, Karas C, Saygili U, Posaci C. 2006. Effects of estrogen, raloxifene, and hormone replacement therapy on serum C-reactive protein and homocysteine levels. Maturitas 53:252–259. Ikonomidis I, Andreotti F, Economou E, Stefanadis C, Toutouzas P, Nihoyannopoulos P. 1999. Increased proinﬂammatory cytokines in patients with chronic stable angina and their reduction by aspirin. Circulation 100:793–798. Jilma B, Dirnberger E, Löscher I, Rumplmayr A, Hildebrandt J, Eichler H, Kapiotis S, Wagner OF. 1997. Menstrual cycle- CRP ACROSS THE MENSTRUAL CYCLE associated changes in blood levels of interleukin-6, a 1 acid glycoprotein, and C-reactive protein. J Lab Clin Med 130:69–75. Kelly RW. 1994. Pregnancy maintenance and parturition: the role of prostaglandin in manipulating the immune and inﬂammatory response. Endo Rev 15:684–706. Kluft C, de Maat MPM. 2001. Determination of the habitual low blood level of C-reactive protein in individuals. Ital Heart J 2:172–180. Kluft C, Gevers Leuven JA, Helmerhorst FM, Krans HMJ. 2002. Pro-inﬂammatory effects of oestrogens during use of oral contraceptives and hormones replacement treatment. Vascul Pharmacol 39:149–154. Kovacs A, Henriksson P, Hamsten A, Wallén H, Björkegren J, Tornvall P. 2005. Hormonal regulation of circulating C-reactive protein in men. Clin Chem 51:911–913. LaMonte MJ, Durstine JL, Yanowitz FG, Lim T, DuBose KD, Davis P, Ainsworth BE. 2002. Cardiorespiratory ﬁtness and C-reactive protein among a tri-ethnic sample of women. Circulation 106:403–406. Libby P, Ridker PM, Maseri M. 2002. Inﬂammation and atherosclerosis. Circulation 105:1135–1143. MacGregor AJ, Gallimore JR, Spector TD, Pepys MB. 2004. Genetic effects on baseline values of C-reactive protein and serum amyloid A protein: a comparison of monozygotic and dizygotic twins. Clin Chem 50:130–134. Macy EM, Hayes TE, Tracy RP. 1997. Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications. Clin Chem 43:52–58. McDade TW, Burhop J, Dohnal J. 2004. High-sensitivity enzyme immunoassay for C-reactive protein in dried blood spots. Clin Chem 50:652–654. McDade TW, Leonard WR, Burhop J, Reyes-Garcia V, Vadez V, Huanca T, Godoy RA. 2005. Predictors of C-reactive protein in Tsimane’ 2 to 15 year-olds in lowland Bolivia. Am J Phys Anthropol 128:906–913. Meier-Ewert HK, Ridker PM, Rifai N, Price N, Dinges DF, Mullington JM. 2001. Absence of diurnal variation of C-reactive protein concentrations in healthy human subjects. Clin Chem 47:426–430. Miller RC, Brindle E, Holman DJ, Shofer J, Klein NA, Soules MR, O’Connor KA. 2004. Comparison of speciﬁc gravity and creatinine for normalizing urinary reproductive hormone concentrations. Clin Chem 50:924–932. Miro F, Aspinall LJ. 2005. The onset of the initial rise in follicle-stimulating hormone during the human menstrual cycle. Human Reprod 20:96–100. Miyamoto N, Mandai M, Suzuma I, Suzuma K, Kobayashi K, Honda Y. 1999. Estrogen protects against cellular inﬁltration by reducing the expression of E-selectin and IL-6 in endotoxin-induced uveitis. J Immunol 163:374–379. Núñez-de la Mora A, Chatterton RT, Choudhury OA, Napolitano DA, Bentley GR. 2007. Childhood conditions inﬂuence adult progesterone levels. PLoS Med 4:e167. Ockene IS, Matthews CE, Rifai N, Ridker PM, Reed G, Stanek E. 2001. Variability and classiﬁcation accuracy of serial highsensitivity C-reactive protein measurements in healthy adults. Clin Chem 47:444–450. O’Connor KA, Brindle E, Holman DJ, Klein NA, Soules MR, Campbell KL, Kohen F, Munro CJ, Shofer JB, Lasley BL, Wood JW. 2003. Urinary estrone conjugate and pregnanediol 3-glucuronide enzyme immunoassays for population research. Clin Chem 49:1139–1148. O’Connor KA, Brindle E, Shofer JB, Miller RC, Klein NA, Soules MR, Campbell KL, Mar C, Handcock MS. 2004. Statistical correction for non-parallelism is a urinary enzyme immunoassay. J Immunoassay Immunochem 25:259–278. O’Connor KA, Brindle E, Miller RC, Shofer JB, Ferrell RJ, Klein NA, Soules MR, Holman DJ, Mansﬁeld PK, Wood JW. 2006. Ovulation detection methods for urinary hormones: precision, daily and intermittent sampling and a combined hierarchical method. Hum Reprod 21:1442–1452. Panter-Brick C, Lunn PG, Baker R, Todd A. 2001. Elevated acute-phase protein in stunted Nepali children reporting low 145 morbidity: different rural and urban proﬁles. Br J Nutr 85: 125–131. Pearson TA, Mensah GA, Alexander RW, Anderson JL, Cannon RO III, Criqui M, Fadl YY, Fortmann SP, Hong Y, Myers GL, Rifai N, Smith Jr SC, Taubert K, Tracy RP, Vinicor F; Centers for Disease Control and Prevention; American Heart Association. 2003. Markers of inﬂammation and cardiovascular disease: application to clinical and public health practice: a statement for healthcare professionals from the Centers for Disease Control and Prevention and the American Heart Association. Circulation 107:499–511. Pepys MB, Hirschﬁeld GM. 2003. C-reactive protein: a critical update. J Clin Invest 111:1805–1812. Pradhan AD, Manson JE, Rifai N, Buring JE, Ridker PM. 2001. C-reactive protein, interleukin-6, and risk of developing type 2 diabetes mellitus. JAMA 286:327–334. Prestwood KM, Unson C, Kulldorff M, Cushman M. 2004. The effect of different doses of micronized 17b-estradiol on Creactive protein, interleukin-6, and lipids in older women. J Gerontol 59:827–832. Ridker PM, Hennekens CH, Rifai N, Buring JE, Manson JE. 1999. Hormone replacement therapy and increased plasma concentration of C-reactive protein. Circulation 100:713– 716. Roberts CW, Walker W, Alexander J. 2001. Sex-associated hormones and immunity to protozoan parasites. Clin Microbiol Rev 14:476–488. Rogers A, Eastell R. 2001. The effect of 17b-estradiol on production of cytokines in cultures of peripheral blood. Bone 29:30– 34. Ropponen A, Aittomäki K, Tikkanen MJ, Ylikorkala O. 2005. Levels of serum C-reactive protein during oral and transdermal estradiol in postmenopausal women with and without a history of intrahepatic cholestasis of pregnancy. J Clin Endocrinol Metab 90:142–146. Ross R. 1999. Atherosclerosis-an inﬂammatory disease. N Engl J Med 340:115–126. Sacks GP, Seyani L, Lavery S, Trew G. 2004. Maternal C-reactive protein levels are raised at 4 weeks gestation. Human Reprod 19:1025–1030. Sattar N, McConnachie A, O’Reilly D, Upton MN, Greer IA, Smith GD, Watt G. 2004. Inverese association between birth weight and C-reactive protein concentrations in the MIDSPAN Family Study. Arterioscler Thromb Vasc Biol 24:583– 587. Schwartz NB. 1974. The role of FSH and LH and of their antibodies on follicle growth and on ovulation. Biol Reprod 10: 236–272. Shell-Duncan B, McDade T. 2004. Use of combined measures from capillary blood to assess iron deﬁciency in rural Kenyan children. J Nutr 134:384–387. Sierksma A, van der Gaag MS, Kluft C, Hendriks HFJ. 2002. Moderate alcohol consumption reduces plasma C-reactive protein and ﬁbrinogen levels; a randomized, diet-controlled intervention study. Eur J Clin Nutr 56:1130–1136. Skouby SO, Gram J, Andersen LF, Sidelmann J, Petersen KR, Jespersen J. 2002. Hormone replacement therapy: estrogen and progestin effects on plasma C-reactive protein concentrations. Am J Obstet Gynecol 186:969–977. Sorensen MV, Leonard WR, Tarskaya LA, Ivanov KI, Snodgrass JJ, Alekseev VP, Krivoshapkin VG, Rifai N. 2006. High-sensitivity C-reactive protein, adiposity, and blood pressure in the Yakut of Siberia. Am J Hum Biol 18:766–775. Tchernof A, Nolan A, Sites CK, Ades PA, Poehlman ET. 2002. Weight loss reduced C-reactive protein levels in obese postmenopausal women. Circulation 105:564–569. Vegeto E, Bonincontro C, Pollio G, Sala A, Viappiani S, Nardi F, Brusadelli A, Viviani B, Cianaand P, Maggi A. 2001. Estrogen prevents the lipopolysaccharide-induced inﬂammatory response in microglia. J Neurosci 21:1809–1818. Vigushin DM, Pepys MB, Hawkins PN. 1993. Metabolic and scintigraphic studies of radioiodinated human C-reactive protein in health and disease. J Clin Invest 91:1351– 1357. American Journal of Physical Anthropology 146 K. WANDER ET AL. Visser M, Bouter LM, McQuillen GM, Wener MH, Harris TB. 1999. Elevated C-reactive protein levels in overweight and obese adults. JAMA 282:2131–2135. Walsh BW, Paul S, Wild RA, Dean RA, Tracy RP, Cox DA, Anderson PW. 2000. The effects of hormone replacement therapy and raloxifene on C-reactive protein and homocysteine in healthy postmenopausal women: a randomized, controlled trial. J Clin Endocrinol Metab 85:214–218. Watts DH, Krohn MA, Wener MH, Eschenbach DA. 1991. C-reactive protein in normal pregnancy. Obstet Gynecol 77:176–180. Wener MH, Daum PR, McQuillan GM. 2000. The inﬂuence of age, sex, and race on the upper reference limit of serum Creactive protein concentration. J Rheumatol 27:2351–2359. American Journal of Physical Anthropology Woodward M, Rumley A, Lowe GDO, Tunstall-Pedoe H. 2003. C-reactive protein: associations with haematological variables, cardiovascular risk factors and prevalent cardiovascular disease. Br J Haematol 122:135–141. Wu T, Mendola P, Buck GM. 2002. Ethnic differences in the presence of secondary sex characteristics and menarche among US girls: the Third National Health and Nutrition Examination Survey, 1988–1994. Pediatrics 110:752–757. Zitzmann M, Erran M, Kamischke A, Simoni M, Nieschlag E. 2005. Endogenous progesterone and the exogenous progestin norethisterone enanthate are associated with a proinﬂammatory proﬁle in healthy men. J Clin Endocrinol Metab 90:6603– 6608.