hormone measures in finger-prick blood spot samples: new...

Hormone Measures in Finger-Prick Blood Spot Samples:New Field Methods for Reproductive Endocrinology

CAROL M. WORTHMAN* AND JOY F. STALLINGSLaboratory for Comparative Human Biology, Department of Anthropology,Emory University, Atlanta, Georgia 30322

KEY WORDS FSH; LH; PRL; testosterone; estradiol; DHEAS;androstenedione; SHBG; puberty; cortisol; saliva; human behavior

ABSTRACT Comparative endocrine studies have notably advanced under-standing of ecological factors that contribute to variation in human reproduc-tive function. Such research has relied on methodological advances thatpermit hormone determinations in samples that are easily and safelycollected, stored, and transported, most recently on measurement of steroidsin saliva. This report seeks to further expand the scope of endocrine researchby demonstrating the value of blood spot samples collected by finger prick. Asa sampling strategy, finger-prick blood spot collection offers the advantages ofshort collection time, low invasiveness, repeatability, absence of postcollectionprocessing, low biohazard risk, and ease of sample storage and transport. Wedocument good sample stability and present sensitive assay methods for arange of steroids and proteins (FSH, LH, PRL, T, E2, DHEAS, androstenedi-one, cortisol, SHGB) in blood spots that require sample volumes of 3–12 µl anddisplay good reliability, specificity, precision, accuracy, and convertibility ofresults to plasma/serum equivalent concentrations. Laboratory evaluationwas augmented by a feasibility study at a remote site in Papua New Guineathat confirmed validity and stability of blood spot collections under fieldconditions. Research applications of blood spot sampling are illustrated with aseries of studies, including cross-sectional surveys for developmental and lifespan endocrinology, a longitudinal, population-based developmental epidemio-logic study of puberty, and serial sampling in a dynamic study of neuroendo-crine response to suckling. We conclude that the sampling features and widerange of measurable biomolecules of blood spots do constitute a methodologi-cal advance for endocrine research. Am J Phys Anthropol 104:1–21, 1997.r 1997 Wiley-Liss, Inc.

During the past two decades, measure-ments of salivary steroids have enhancedendocrine research (Malamud and Tabak,1993), broadening the scope of study topreviously inaccessible populations andstudy sites, such as remote field locations,homes and work places, exercise facilities,and educational settings. The advantages ofsalivary assessments of hormones over bloodstem primarily from the ease of samplecollection, providing a completely painlesssampling technique that is minimally disrup-tive to the daily routine and highly accept-

able to most subjects. Steroid hormones are,moreover, quite stable in unchilled, pre-served saliva specimens, which eases stor-age constraints.

Despite these manifest advantages, sali-vary measures have certain drawbacks. In

Contract grant sponsor: W.T. Grant Foundation; contract grantnumber 92-1489-92, 94-1489-2; Contract grant sponsor: Univer-sity Research Committee of Emory University; Contract grantsponsor: NIMH; contract grant number MH48085.

*Correspondence to: Carol M. Worthman, Department of An-thropology, Emory University, Atlanta, GA 30322.

Received 12 September 1996; accepted 11 June 1997.

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 104:1–21 (1997)

r 1997 WILEY-LISS, INC.

fact, salivary samples present several logis-tical and physiological limitations. Samplevials are bulky, breakable, potentially leaky,and unwieldy to transport; such consider-ations are most significant for field workersat remote sites who must transport salivavials long distances in a range of environ-mental circumstances (terrain, climate, alti-tude) using diverse means of conveyance(e.g., foot, air, or llama). More importantscientifically is the fact that saliva does notreflect circulating concentrations of peptideor protein hormones, binding proteins, conju-gated steroids, or other membrane-insolubleconstituents. Saliva represents the free ste-roid fractions rather than that bound tocarrying proteins and thus is thought torepresent the bioavailable moiety. While thishas been viewed as an advantage, salivarystudies provide limited estimates of totaloutput by peripheral glands (gonads, adre-nals) and preclude evaluation of centralneuroendocrine regulation. Further, thatmerely 2–8% of circulating concentrationsare present in saliva places severe technicaldemands on assays for steroids of low concen-trations (e.g., estradiol) and prevents theiruse in developmental studies.

In 1990, faced with these limitations, wecommenced development of new methodsthat would be similar to salivary collectionsin noninvasiveness and acceptability butthat would obviate some of the problems andlimitations we had encountered by relyingsolely on salivary methodology when bloodsamples from venipuncture were impos-sible. The ability to study central neuroendo-crine regulation was crucial to our cross-population investigations of developmentalendocrinology and factors influencing matu-rational timing and sociobehavioral and en-vironmental effects on adult reproductivefunction and stress. Since most of our stud-ies involved remote field settings, an impor-tant criterion for any new method was sam-ple stability to allow some sample storage atthe site since immediate transportation tothe laboratory was rarely possible. Based onpublished methods for blood spot thyroidstimulating hormone (TSH) and their wide-spread use in neonatal screening (Brom-bacher et al., 1988; Torresani and Scherz,1986; Waite et al., 1987), for blood spot

prolactin (PRL) assays (Bassett et al., 1986),and for some blood spot steroids (Egan et al.,1989; Hofman et al., 1985; Kraiem et al.,1980; Petsos et al., 1986; Thomson et al.,1989), the mode of sampling we chose wasblood spot by finger prick.

After 6 years of effort in method develop-ment, we now have blood spot radioimmuno-assays (RIAs) and fluoroimmunometric as-says (FIAs) for determination of gonadal(testosterone (T), estradiol [E2]) and adre-nal (androstenedione (A), dehydroepiandros-terone-sulfate (DHEA-S), and cortisol [C])steroids, pituitary hormones (follicle stimu-lating hormone (FSH), luteinizing hormone(LH), and PRL), and the primary sex hor-mone carrying protein (sex hormone bindingglobulin [SHBG]). These methods have metour sample collection and handling criteriaas well as our scientific interests in achiev-ing endocrine profiles of both steroid andprotein hormones from a single collection ina field setting, interests heretofore met onlyby serum or plasma samples from venipunc-ture.

Although blood collection is inappropriatefor some populations, these methods providea means by which blood samples can beobtained where the major concerns havebeen invasiveness, storage, and handling.The purpose of this paper is to provide otherresearchers with methodological informa-tion and examples of studies performed inour laboratory that demonstrate their appli-cation. Our own experience suggests thatthese methods will prove useful not only foranthropologists but also for a wide range ofepidemiological, clinical, psychobiological,and interdisciplinary research. They further-more expand the range and power of endo-crine investigation of normal individuals ineveryday settings.

COMPARISON OF BLOOD SPOTSWITH SALIVA AND SERUM

Collections of finger-prick samples areminimally invasive since the devices usedare designed for repeated collections by dia-betics for glucose self-monitoring. The as-says themselves require very little blood; infact, all of the aforementioned analytes canbe measured simultaneously from collectionof four drops of blood (200 µl or four circles

2 C.M. WORTHMAN AND J.F. STALLINGS

about the size of a dime). Unlike saliva,sample volume for blood spots is low becausehormone levels are generally one or twoorders of magnitude higher in blood than insaliva. Furthermore, advances in assay sys-tems yield ever more sensitive measures,particularly for proteins, requiring concomi-tantly diminished sample volumes. Withhighly sensitive RIAs and FIAs (Haavisto etal., 1990; Lovgren et al., 1985; Stenman etal., 1985), measurements can be performedon very small sample sizes. Although salivacollection is totally painless, up to 15 min ormore of subject time is needed to collect the 5ml volumes often required for analysis. Sub-ject collection time for finger pricks, on theother hand, is more on the order of a coupleof minutes.

One reccurring problem we and othershave experienced with saliva collections isblood contamination, a highly prevalentproblem in field settings where diet type andoral hygiene may conduce to bleeding gums(Campbell, 1994). Because of the large blood-saliva disparity in hormone concentrations,even minute amounts of blood contamina-tion can cause highly inaccurate results.Blood in saliva can be detected by tech-niques such as dipsticks used to identifyblood in urine (VWR Scientific Products,Stone Mountain, GA) so that contaminatedsamples can be discarded. This strategem isplagued by false positives (Beall et al., 1992)and can jeopardize the study by reducing thesample size and eliminating important, irre-placable samples. Others have reported atechnique that may be useful in separatingblood-source hormones from saliva (Camp-bell, 1994), but these techniques increaseanalytical time and cost.

Fewer factors compromise the validity ofblood spot samples for analysis. Blood fromthe finger prick is dropped directly ontospecially designed filter paper (#903 Schlei-cher & Schuell (S&S), Keene, NH) thatabsorbs the blood uniformly to make a homo-geneous, evenly dispersed blood spot that isessential for accurate analysis (NCCLS,1992). Sample collection errors that cangenerate unsuitable samples include 1) in-complete absorption, which occurs when aninadequate amount of blood is smeared onthe paper surface (see details of blood collec-

tion that follow), 2) nonuniform samples,when one blood drop is applied on top ofanother, and 3) exposure of samples to directsources of heat such as sunlight (or to blood-eating insects!).

Since release of many hormones is pulsa-tile (Dunkel et al., 1990; Marshall et al.,1991; Stenman et al., 1985; Wu et al., 1991),one limitation of serum/plasma samples hasbeen that a single-point sample may betaken during the peak of the secretory pulseand thus fail to represent tonic or meanconcentrations (Campbell, 1994; Ellison,1988). The invasiveness of venipunctures orcatheterization in field settings have left fewalternatives to this problem. Finger pricks,on the other hand, allow multiple collec-tions; we have used two collections 20 minapart in our epidemiologic studies of theassociations between pubertal developmentand mental health among US adolescents(Angold et al., 1995) and three serial collec-tions 20–25 min apart in our investigationsof lactational infertility among nursingNepali women (Stallings et al., 1996). Fin-ger pricks also allow self-sampling; in ourstudy of attitudinal and hormonal changesin expectant fathers (Worthman and Stall-ings 1994; Worthman et al., 1991), partici-pants collected blood spot samples thriceweekly for 9 weeks, storing samples in theirhome refrigerator until weekly pickup byinvestigators.

Steroids in saliva reflect the portion ofcirculating hormone that is able to diffuseacross acinar cells lining the salivary gland(Riad-Fahmy et al., 1987; Vining and McGin-ley, 1987). Since this process is analogous tothe diffusion of steroids across cell mem-branes of target cells, a reported advantageof salivary over plasma hormone levels isthat the former reflect the biologically active(the free or unbound) portion available totarget cells. Indeed, several reports docu-ment strong correlations between salivarysteroid concentrations and target tissue func-tion (Beall et al., 1992; Ruutiainen et al.,1987; Osredkar et al., 1989).

Blood hormone assays reflect both boundand free forms, and several studies haveyielded strong correlations between totalplasma and salivary free steroid levels (Rill-ing et al., 1996; Sannikka et al., 1983; Wang

3BLOOD SPOT HORMONE MEASURES

et al., 1981; reviewed in Malamud and Ta-bek, 1993). For example, in a recent study of218 Zimbabwean male adolescents 11–23years of age (Rilling et al., 1996), we found ahigh correlation between matched salivaryand blood spot T levels (r 5 .83) and ex-pected increases in both salivary and bloodspot T across pubertal stage (assessed byself-reported Tanner genital stage ratings).Although blood spot T levels were slightlymore effective than salivary levels in distin-guishing the stage of genital development,the difference was not sufficient to establishthe superiority of one measure over theother as an index of the bioavailability ofcirculating plasma T.

SHBG is the primary carrying protein ofboth T and E2 and the most significantdeterminant of the proportion of free steroidin circulation (Anderson, 1974; Dunn et al.,1981; Pardridge, 1981; Petra, 1991; Siiteriet al., 1982). Certain physiological states(e.g., pregnancy, obesity, puberty, and in-creased energy expenditure) and pathologi-cal conditions (e.g., cirrhosis of the liver andCushing’s disease) alter SHBG and thus theratio of free vs. bound steroids (Anderson,1974; Blank et al., 1978; Carter et al., 1983).Under such conditions, blood spot measuresof both SHBG and sex steroid are useful fordetermining a ‘‘free index,’’ an indicator ofthe biologically available steroid portion, bycalculating a ratio between the total steroid

level and SHBG. Such measures also allowinvestigation of whether individual andpopulation variation in SHBG may contrib-ute to variation in endocrine function andregulation.

One important advantage of salivasamples over serum and plasma has beenstability under field conditions. Ellison(1988) determined that steroids in salivatreated with 0.01% sodium azide and left atroom temperature for as long as 6 monthsshowed no significant change in concentra-tions. By contrast, serum and plasmasamples must be frozen within a 24 h periodto insure stability for most analytes. There-fore, for field collections, the substantialtime, labor, and equipment involved in bloodsample processing, storage, and transport(on dry ice or in a cryogenic freezer, assum-ing sources of processed gasses are availablenear the site) deter more frequent use ofblood samples.

The hormones we have measured in bloodspots, although not stable as long as salivaat room temperature, remain stable for up to3 weeks as long as they are kept dry and outof direct sunlight. In humid environments,storage with dessicant prevents mildew ofthe samples. Refrigerated samples (4°C) arestable for several weeks (see Table 1 forhormone-specific stability results) and canbe removed for transport at room tempera-ture without degradation. Freezer-stored

TABLE 1. Source of reagents, type of assay, range of blood spot standards, and duration of samplestability by temperature1

Hormone Kit manufacturer Assay Range of standards

Sample stability

Room temperature 4°C 37°C

Pituitary hormonesFSH2 Wallac, Inc. FIA 0.5–128 IU/L 8 weeks 8 weeks 1 weekLH3 Wallac, Inc. FIA 0.75–125 IU/L 8 weeks 8 weeks 1 weekPRL4 Wallac, Inc. FIA 1.25–125 ng/mL 3 weeks 8 weeks 1 week

Gonadal hormonesT Binax (South Portland, ME) RIA 6–1,000 ng/dL 3 weeks 8 weeks 5 daysE2 Pantex (Santa Monica, CA) RIA 5–1,500 pg/mL 3 weeks 8 weeks NS

Adrenal hormonesDHEA-S Pantex RIA 25–4,000 ng/mL 4 weeks 8 weeks 1 weekA DSL (Webster, TX) RIA 0.05–5 ng/mL 4 weeks 8 weeks 2 weeksC Pantex RIA 0.5–16 µg/dL 4 weeks 8 weeks NS

Binding proteinSHBG Wallac, Inc. FIA 3.125–100 nmol/L 2 weeks 8 weeks NS

1 A, androstenedione; C, cortisol; DHEA-S, dehydroepiandrosterone sulfate; E2, estradiol; FSH, follicle stimulating hormone; LH,luteinizing hormone; NS, not stable; PRL, prolactin; SHBG, sex hormone binding globulin; T, testosterone.2 Calibrated against the second IRP (78/549).3 Calibrated against the WHO second International Standard (80/552).4 Calibrated against the WHO third International Standard (84/500).


samples are stable for at least a year. Sincesamples are dry, the possibility of spills iseliminated, and papers can simply be car-ried in zip-lock bags, making transport ofblood spot samples considerably easier thantransporting saliva or serum. Moreover, dry-ing of samples destroys HIV and hepatitisviruses and thus substantially reduces bio-hazard (Knudsen et al., 1993; Meredith andHannon, 1993).

A final advantage of serum measures oversaliva is the enormous literature availablebased on serum/plasma assays. Blood spotvalues are highly correlated with plasmavalues and thus can be converted to plasmaequivalents using linear regression coeffi-cients (Table 4), providing researchers withthe advantage of comparability to this richendocrine literature.

METHODOLOGYSample collection

The materials needed for collection offinger pricks are minimal. Samples are col-lected on filter papers (#903; S&S) highlystandardized to absorb blood in a homoge-neous manner so that uniform punches fromany section of the sample will yield the samequantity of blood (NCCLS, 1992). After thefingertip is wiped with an alcohol swab, thefinger is pricked with a lancet (Unilet BloodLancets; VWR Scientific Products, StoneMountain, GA) fitted into a spring-type autolancet device (Autolet Lancet Mark II; VWRScientific Products) designed to minimizepain to the subject. In fact, this device isused by diabetics for chronic self-samplingfor blood glucose monitoring. Where handsare calloused or where investigators havelittle experience in finger-prick collection,very slightly larger lancets [MicrotainerSafety Flow Lancet; VWR Scientific Prod-ucts] should be used. Whichever device isemployed, it must be pressed firmly to thefinger for adequate penetration. When bloodflow begins, the first bit is wiped away with atissue since this drop may contain tissuefluids. The next drops are placed on each offive preprinted circles on the S&S paper thathas been previously labeled with appropri-ately identifying subject information. Thesample is then left at room temperature toair-dry, usually 3–4 h, depending on humid-

ity levels. After drying, the card is placed ina zip-lock bag and stored at room (22°C) orrefrigerator temperature (4°C) for the timesspecified in Table 1 or frozen (220°C) untilshipment to the laboratory for analysis.

Moderately experienced investigators canreliably obtain five drops per puncture. Thegreatest potential source of error in samplecollection comes from application of the sam-ple to the paper. The paper is designed towick away the blood drop without the fingeractually touching the paper. In other words,the finger is held slightly above the paperand the blood is simply dropped onto theprinted circle, penetrating both sides of thepaper. In cold climates, care should be takennot to allow the samples to freeze and thaw,while in warm climates samples should bekept away from direct sunlight and in arelatively cool area (i.e., by placing them inthe shade). During rainy seasons or highlyhumid conditions, dessicant should be usedto avoid mildew. Papers should not be com-pressed before use, because compression mayaffect the papers’ absorbent properties. Aword of caution concerning insects: one re-searcher had the unfortunate experience ofhaving exposed samples eaten by flies whileair-drying, so placing them in a cardboardbox while drying may be prudent.

Blood spot hormone assays

The following section is provided primar-ily for readers familiar with RIAs and FIAswho are interested in incorporating themethods in their own laboratories or whoplan to collaborate with other laboratoryscientists.

Preparation of blood spot controls andstandards. Commercial control sera arepurchased from Bio-Rad, ECS Division (Ana-heim, CA), and standards are supplied bythe kit manufacturers. Controls are pre-pared according to the manufacturer’s in-structions by adding 5 ml deionized water tothe lyophilized product, allowing it to standat least 15 min to reconstitute, and theninverting it several times to thoroughly mixthe contents. Standards are supplied in aserum-based liquid form so reconstitution isnot necessary.


Blood spot standards and controls areprepared by mixing reconstituted controlsand standards 1:2 with red blood cells thathave been washed three times with normalsaline. For some assays, whole blood poolsfrom EDTA-anticoagulated venipuncture col-lections are also used as controls. Five ali-quots (50 µl each) of prepared blood stan-dards and controls are dropped onto thesample collection papers, allowed to dry atroom temperature overnight, and then storedin an air-tight container at 220°C. Themanufacturer provides expected values forthe commercial controls and blood spot val-ues for controls, and standards are half thatexpected for serum or plasma since they arediluted 1:2 with red blood cells. Blood spotcontrols and standards are stable for at least1 year when stored frozen at 220°C.

Principles of the assays. The blood spotPRL, FSH, LH, and SHBG assays are modi-fications of commercially available FIA kitsfor measurement of these hormones in se-rum or plasma (DELFIA; Wallac, Inc., Gai-thersburg, MD). The principle involves thedirect determination of hormonal levels byusing two monoclonal antibodies directedagainst separate hormonal antigenic sites,one antibody immobilized on microtitre wellsand the other europium-labeled. Europiumions are dissociated from the antibody com-lex, which then fluoresces in conjunction with achelating agent. Shaking and washing of assaystrips and time-resolved fluorimmetry are per-

formed by instruments purchased from the kitmanufacturer (Wallac, Inc.).

The blood spot T, E2, A, DHEA-S, and Cassays are modifications of commerciallyavailable serum/plasma RIA kits (see Table1 for manufacturers). The basic principle iscompetition between a radioactive and non-radioactive antigen for a fixed number ofantibody sites (Yalow and Berson, 1971).The concentration of hormone in the sampleis inversely proportional to the amount of125I-labeled hormone bound to the antibody.The free and bound antigens are separatedby the addition of a second antibody–polyeth-ylene glycol (PEG) reagent followed by cen-trifugation. After the supernatant is de-canted, the amount of radioactivity in theantibody bound tracer complex is measuredin a gamma counter (Cobra Gamma Counter,model 5005; Packard Instrument, Co., Down-ers Grove, IL).

Preparation of reagents. For all FIAs,europium-labeled antibodies are diluted withbuffer supplied in the kits. Dilutions ofantibodies and separating reagents for theRIAs are made with Dulbecco’s phosphate-buffered saline (Gibco, Grand Island, NY), pH7.4, containing 0.1% gelatin. This workingbuffer is prepared by adding 100 mg of gelatinto 100 ml of Dulbecco’s buffer and heating to45°C to dissolve. Table 2 shows the dilutionsmade for the reagents in each assay. All otherreagents required for each assay are suppliedin the respective kit in ready-to-use form.

TABLE 2. Number of blood spot punches and volume of buffer used in elution of blood spotsand dilutions of kit reagents1

Assay

Elution of Standards, Controls, Samples

Dilution ofantibody2

Incubationtime (min)with tracer

Dilution ofsecond antibody

Number ofpunches per well

Assay buffervolume (µL)

Fluoroimmunometric assaysFSH 1 200 (1:100) 30LH 1 200 (1:150) 15PRL 1 200 (1:75) 90SHBG 1 500 (1:50) 120

RadioimmunoassaysT 4 300 (1:4) (1:2)E2 4 200 (1:8) (1:4)DHEA-S 1 75 ND NDA 4 200 (1:6) (1:2)C 1 100 (1:4) (1:4)

1 ND, no dilution.2 FIA: europium-antibody solution; RIA: first antibody


Elution of blood standards, controls,and samplesFluoroimmunometric assays

FSH, LH, and PRL assays. Blood iseluted from the sample paper using kitassay buffer. The assay strips, consisting of12 microtitre antibody-coated wells, arerinsed once with wash solution using anautomatic washer (Platewash automaticwasher, model 1296-024; Wallac, Inc.). Bloodspot standards, controls, and samples areremoved from the freezer, and duplicate 2.5mm punches (using a hole punch availableat local office supply stores), equivalent to 3µl whole blood, are transferred to the wells(one punch per well) with tweezers. Within30 min of placing the spots, 200 µl assaybuffer is delivered to each well. With anautomatic plateshaker (Plateshake auto-matic shaker, model MPS-4; Wallac, Inc),the plates (consisting of eight microtitrestrips) are rotated at 150 rpm for 10 min atroom temperature. The wells are then placedin an air tight container in the refrigerator(4°C) overnight.

SHBG assay. Blood spot standards,controls, and samples are removed from thefreezer, and a 2.5 mm punch from each istransferred to a 12 3 75 mm glass tube. Kitassay buffer (500 µl) is added to the tubes;the tubes are shaken on the automaticshaker for 10 min at 50 rpm at room tempera-ture, covered tightly with parafilm, and thenincubated overnight at 4°C.

Radioimmunoassays. Blood is eluted fromthe sample paper using working buffer.Buffer volume and number of punches varywith each assay and are listed in Table 2.Whole blood sample volumes per well rangefrom 3–12 µl. The specified number of 2.5mm punches is transferred to labeled 12 375 mm glass tubes, and a volume of workingbuffer (see Table 2) is added to each tube.The tubes are covered with parafilm androtated on the automatic shaker at 50 rpmfor 1 h at room temperature. After shaking,the tubes are placed in a 4°C refrigeratorovernight.

Assay proceduresFluoroimmunometric assays

FSH, LH, and PRL assays. Follow-ing overnight incubation, the wells are re-moved from the refrigerator and placed onthe automatic shaker at 50 rpm for 1 h. Thedilution of europium-labeled tracer solutionis prepared as specified in Table 2 using kitassay buffer. Following the 1 h shaking, thefilter paper discs are removed from the wellsusing vacuum aspiration with a pasteurpipette. The strips are washed twice with kitwash buffer using the automatic washer,and 200 µl of tracer solution is added to thewells. The strips are then placed on theautomatic shaker at 150 rpm for 2 min andthen at 50 rpm for the time specified in Table2. Each strip is washed six times using theautomatic platewasher, 200 µl of kit enhance-ment solution is added to each well, and thestrips are rotated on the automatic shakerat 50 rpm for 5 min. After shaking, the wellsare incubated at room temperature for 10min, and the fluorescence is measured in thefluorometer (Arcus time-resolved fluorom-eter, model 1230; Wallac, Inc.). Concentra-tions are interpolated from the standardcurve using a linear/log data reductionmethod and converted to plasma equiva-lents derived from regression analysis ofmatched blood spot/plasma samples (seeTable 4 for regression coefficients for eachassay). All data analyses are performed us-ing RIA-Smart and Expert QC Software (Pack-ard Instrument Co., Downers Grove, IL).

SHBG assay. Following overnight in-cubation, the tubes are removed from therefrigerator and placed on the automaticshaker at 50 rpm for 1 h. The assay strips(12 wells per strip) are rinsed once with kitwash solution using the automatic washer. A25 µl aliquot of each eluate is transferred toduplicate wells, and 100 µl of assay buffer isadded. The strips are shaken on the auto-matic shaker at 50 rpm for 2 h and thenwashed twice with wash solution. The dilu-tion of europium-labeled tracer solution isprepared as specified in Table 2, and 100 µlis added to each well. The strips are shakenfor 2 h at 50 rpm and then washed six times


with wash solution using the automaticwasher. Enhancement solution (200 µl) isadded to each well, and the wells are shakenon the automatic shaker at 50 rpm for 5 minincubation. Following shaking, the wells areincubated at room temperature for 10 min,and the fluorescence is measured in thefluorometer. Concentrations are interpo-lated from the standard curve using a linear/log data reduction method and converted toserum equivalents derived from regressionanalysis of matched blood spot/serumsamples (see Table 5 for regression coeffi-cients). Note that EDTA-anticoagulatedblood cannot be used in this SHBG assay;therefore, reported comparisons of SHBGmeasured in whole and processed blood arebased on matched blood spot and serumsamples.

Radioimmunoassays. Following overnightincubation, the tubes are removed from therefrigerator and shaken on the automaticshaker at 50 rpm for 1 h at room tempera-ture. Specific dilutions of reagents, volumes,incubation temperatures, and times aregiven in Tables 2 and 3. Following the 1 hincubation, a volume of the blood spot elu-ate, 125I-labeled hormone, and diluted firstantibody are added to duplicate 12 3 75 mmpolypropylene tubes. The tubes are incu-bated, and then the antibody-bound and freeantigen are separated using a second anti-body/PEG reagent and centrifugation (Sor-vall RC3C Centrifuge; Dupont, Wilmington,DE). After the supernatant containing un-bound antigen is decanted, the amount ofradioactivity in the pellet (containing the

bound portion) is measured in the gammacounter. Concentrations are interpolatedfrom the standard curve using a spline datareduction method and converted to plasmaequivalents derived from regression analy-sis of matched blood spot/plasma samples(see Table 4 for regression coefficients foreach RIA). All data analyses are performedusing RIA-Smart and Expert QC Software.

Assay performance

Sensitivity. The sensitivity of the FIAs isdefined as the dose required to enhancefluorescing two standard deviations (SD)above zero dose, and the sensitivity of theRIAs is defined as the quantity of unlabeledhormone required to inhibit binding of tracerby an amount equal to 2 SD below the meanbinding observed in the absence of unla-beled hormone. Sensitivity doses calculatedfor each FIA and RIA are listed in Table 5.

TABLE 3. Assay-specific volumes (mL) of blood spot eluates and reagents, incubation and centrifugation times(min) and temperatures (°C) for each radioimmunoassay1

Protocol step Specification T E2 DHEA-S A C

Pipet Blood spot eluate 100 75 502 75 40Tracer 50 20 100 50 30First antibody 100 500 500 100 100

First antibody incubation TimeTemperature (°C)

ONRT

ONRT

30RT

4037

3037

Pipet Second antibody reagent 500 500 500 500 500Second antibody incubation Time

Temperature20

RT60

RT10

RT15

RT10

RTCentrifugation g 2,300 2,300 2,300 2,300 2,300

Time 60 60 30 20 60Temperature 4 22 22 22 22

1 ON, overnight (18–24 h); RT, room temperature (20–24°C).2 A further 1:10 dilution of the blood spot eluate is made with kit buffer (50 µL eluate plus 500 µL kit diluent); 50 µL of this dilution isused in the assay.

TABLE 4. Comparisons of matched blood spot andplasma/serum hormone measures: Pearson correlation

coefficients (r), number of comparisons (n), r-square,and regression coefficients from simple linear regressionanalysis used to calculate plasma/serum equivalents1

Assay n r r-square Regression formula

FSH 39 0.980 0.960 Y 5 0.42 1 2.2(X)LH 39 0.977 0.956 Y 5 0.07 1 1.9(X)PRL 35 0.985 0.970 Y 5 0.73 1 2.0(X)SHBG 20 0.934 0.873 Y 5 210.0 1 2.0(X)T 56 0.979 0.958 Y 5 0.07 1 1.4(X)E2 21 0.984 0.968 Y 5 14.7 1 1.6(X)DHEA-S 89 0.990 0.980 Y 5 8.17 1 2.2(X)A 51 0.978 0.956 Y 5 0.03 1 1.4(X)C 21 0.926 0.857 Y 5 1.99 1 2.0(X)1 X, blood spot hormone concentration; Y, plasma/serum hormoneconcentration.


Precision. Precision was evaluated usingblood spot controls and calculating intraas-say coefficients of variation (CV) from concen-trations of multiple samples assayed in asingle assay and interassay CVs from concen-trations determined for blood spot controlsfrom multiple assays. Results for each assayare given in Table 5.

Linearity. Linearity was evaluated by seri-ally diluting a high plasma sample with asample of stripped human plasma. Aliquotsof the high sample and dilutions were mixed1:2 with washed red cells. A 50 µl aliquotwas dropped onto the S&S paper, allowed todry at room temperature overnight, andthen stored in an airtight container at 220°Cuntil assayed. For each assay, observed val-ues and percentages of expected values foreach dilution are reported in Table 5.

Accuracy, recovery. Varying concentrationsof hormone were added to plasma samplesoriginally containing lower and higher lev-els of endogenous hormone. Aliquots weremixed 1:2 with washed red cells and appliedto the S&S paper as previously described.Average recoveries, calculated as percent ofexpected values, are listed in Table 5 foreach assay.

Stability. For all assays, standards andcontrols that have been stored frozen at220°C for at least one full year fall within a10% CV range of the values determinedimmediately after preparation. Table 1 showsthe length of time that samples remainwithin a 10% CV range of their initial valueafter having been stored for 1 week intervalsat 4°C, room temperature, and 37°C for a

TABLE 5. Assay performance characteristics

Fluoroimmunometricassays FSH LH PRL SHBG

Sensitivity 0.13IU/L 0.26IU/L 0.11ng/mL 0.2nmol/LAverage % recovery 89.7 100.7 102.8 104.1Intraassay % CV for Bio-Rad

controlsLow 7.8 10.6 3.0 13.2Medium 5.3 7.7 10.5 1

High 9.9 3.5 4.9 1

Interassay % CV for Bio-Radcontrols

Low 9.2 11.6 5.9 14.5Medium 8.6 7.2 8 1

High 5.9 7.8 7.8 1

Linearity2

(1:1) 26.1 IU/L 15.4 IU/L 28.4 ng/mL 61.5nmol/L(1:2) 13.7 IU/L(105) 7.7 IU/L(100) 15.2 ng/mL(107) 31.2nmol/L(101)(1:4) 7.4 IU/L(113) 3.95IU/L(103) 7.5 ng/mL(105) 12.6nmol/L (82)

Radioimmunoassays T E2 DHEA-S A C

Sensitivity 6.3 ng/dL 9 pg/mL 8.0 ng/mL 0.012 ng/mL 0.46 µg/dLAverage % recovery 100.7 103.4 98.6 105.5 94.6Intraassay % CV for

Bio-Rad controlsLow 7.6 14.4 8.6 10.3 11.9Medium 8.3 7.4 6.5 7.9 5.7High 7 4 7.5 10.3 9.3

Interassay % CV forBio-Rad controls

Low 13.9 15 11.6 11.1 12.9Medium 12.3 11.6 5.9 11.3 8.5High 11.8 8.4 10.6 9.8 6.2

Linearity2

(1:1) 292 ng/dL 170.9 pg/mL 457 ng/mL 0.91 mg/mL 21.6 µg/dL(1:2) 130 ng/dL (89) 87.9 pg/mL (103) 229 ng/mL (100) 0.49 ng/mL (108) 11.6 ng/mL (107)(1:4) 72 ng/dL (99) 42.0 pg/mL (98) 104 ng/mL (91) 0.21 ng/mL (92) 5.8 ng/mL (107)

1 Concentrations same as low Bio-Rad.2 Numbers in parentheses are [(observed value 3 serial dilution factor)/expected value] 3 100.


period of up to 8 weeks prior to freezing at220°C.

Comparability to plasma samples. Wholebloods from venipunctures were spotted ontoS&S papers, while portions were centri-fuged and the plasmas (EDTA-anticoagu-lated) or serum (for the SHBG assay) with-drawn. Samples were evaluated for directplasma/serum and blood spot measures.Table 4 shows Pearson correlation coeffi-cients, r-square values, and the regressioncoefficients derived from simple linear re-gression analyses used to calculate plasma/serum equivalents from blood spot hormoneconcentrations.

MEASURING HORMONES FROM BLOODSPOTS COLLECTED IN REMOTEFIELD SETTINGS: A FEASIBILITY

STUDY AMONG THE HAGAHAIOF PAPUA NEW GUINEA

The Hagahai are a recently contacted(1984) forager-swiddenist population(n , 295) living on the northwest edge of theSchrader Range in Papua New Guinea (Jen-kins et al., 1989). The lateness of contactreflects the remoteness of the site, at analtitude of 350–2400 m and accessible onlyby helicopter or by 1 week’s walk from thenearest government post. In conjunction withregular health exams, demographic surveys,and other ongoing biomedical and ethno-graphic research headed by Carol Jenkinswith the Papua New Guinea Institute forMedical Research (IMR), endocrine mea-sures (from plasmas collected by venipunc-tures) and growth data (heights, weights,and skin folds) have been collected over a 10year period to assess health status andacculturative and ecological impact on juve-nile growth and development. Given theremoteness of the area, the logistical difficul-ties of sample collection and transport, andthe technical field support of the IMR, theHagahai site was an ideal location to test thefeasibility of blood spot hormonal measures.Because the ongoing health surveillance pro-tocol involved venipuncture, matched plasmaand blood spot samples could be collected forcomparative endocrine analysis. We had pre-viously established that there was no differ-ence in gonadotropin values assayed from

blood spots produced by dropping wholeblood from a finger prick vs. those producedwith whole blood from a venipuncture (pairedt-test of difference between finger-prick andvenipuncture blood spot hormonal mea-sures: P .17 for FSH, P .25 for LH). Immedi-ate application of blood drops from the veni-puncture onto the S&S papers prior tocentrifugation of the sample therefore al-lowed production of matched plasma (EDTA-anticoagulated) and blood spot samples with-out any extra blood sampling.

On the day prior to departure for the field(accessed in this case by a 1 hour helicopterflight from Mt. Hagen), blood spot samplesfrom the two authors and blood spot controlsfrom commercial control products were pre-pared as described above at the laboratoriesof the IMR at Goroka, PNG. One set ofcontrols and samples was dried overnightand then stored at 220°C. The other set ofblood spots was taken to the field site andexposed to the same environmental condi-tions as the Hagahai samples that werecollected over the next 8 days.

Field collection of samples proceeded asfollows. Immediately following venipunc-ture, blood from the syringe was droppedonto the prelabeled S&S papers, and theremaining sample was placed in an EDTAcollection tube. After centrifugation (with agenerator-powered centrifuge), plasma wastransferred to cryogenic vials and frozenimmediately in a liquid nitrogen cryogenicfreezer. Blood spot samples, on the otherhand, were allowed to air-dry (approxi-mately 4 h) and then placed in zip-lock bagscontaining dessicant and stored for an 8 dayperiod at environmental temperatures (rang-ing from 22–28°C). At the end of the 8 daysampling period, all Hagahai and controlsamples were taken to the IMR and storedat 220°C. Approximately 2 months aftercollection, plasma and blood spot samplesand controls were shipped frozen to theLaboratory for Comparative Human Biol-ogy, Emory University, where they werestored frozen at 220°C until assay. Pitu-itary, adrenal, and gonadal hormones (LH,A, and T) were measured in both blood spotsand plasmas from 27 males and 14 females,9–25 years of age. Excluded from statisticalanalysis are LH and T for nine boys whose


values fell below the sensitivity limit for theblood spot and plasma assays and resultsfrom two pregnant females (LH levels ex-ceeding the standard curve range of 125IU/L).

Table 6 shows expected ranges for bloodspot controls and hormone measures fromcontrols stored frozen at the IMR from timeof preparation and those that were taken tothe field and exposed to environmental con-ditions. Means and standard errors (SE) forboth plasma and plasma-equivalent bloodspot values, Pearson correlation coefficients(r), r-squares, and P values from simplelinear regressions as well as paired t-test Pvalues evaluating differences in plasma andplasma-equivalent blood spot values areshown in Table 7.

Assay results for all blood spot commer-cial controls—those kept frozen at the IMRand those that traveled to the field site—fellwithin the expected ranges. Similarly, hor-mone values for the sample aliquots storedat the IMR differed little from results fromthose taken to the field and treated identi-cally to the samples collected at the fieldsite. Further, strong correlations of matched

plasma and blood spot samples, as shown inTable 7, suggest that blood spot samplesrepresented as adequate a sampling methodas the more cumbersome, costly, time-consuming, and invasive venipunctureplasma samples. Such results support ourcontention that blood spot sampling is auseful as well as practical field method forstudying central and peripheral endocrineregulators of reproductive maturation andfunction.

RESEARCH APPLICATIONS OF BLOODSPOT METHODS

In this section, we briefly describe a set ofstudies which illustrate the practical andscientific value of blood spot sampling. Thelogistical advantages discussed above allowa broader scope of endocrine research includ-ing more cross-cultural and population-based epidemiological studies on a variety oftopics and research questions. Our experi-ence has shown that large numbers ofsamples can be collected over short periodsof time even in remote settings, that largelongitudinal off-site epidemiological studies(even child studies) can be undertaken, thatrepeat samplings of both children and adultsare possible, and that well-motivated and-instructed subjects can self-sample and mailsample papers directly to the lab for storageand analysis.

Cross-sectional developmental study

Blood spots offer advantages for develop-mental studies because they allow mini-mally invasive measurement, acceptable tochild and parent, of the very low levels ofregulatory protein and peripheral steroidhormones present through early puberty.Assay of these hormones is required fordetecting and tracking puberty and provides

TABLE 6. Manufacturer’s expected ranges (MER) for blood spot commercial controls compared to blood spothormone values for samples stored frozen at the Institute for Medical Research, Goroka, Papua New Guinea (IMR)

and to samples exposed to field conditions at the Hagahai site

Bio-Radcommercial

controls

A (ng/mL) LH (IU/L) T (ng/dL)

MER IMR Field MER IMR Field MER IMR Field

Level I 0.10–0.44 0.32 0.35 1.3–2.2 1.7 1.9 33.2–58.8 47.4 50.2Level II 1.11–1.72 1.36 1.55 7.5–11.2 8.3 8.4 499–645 583 501Level III 3.70–5.89 3.87 4.19 21.2–31.5 26 24.1 1 1 1

1 Concentrations exceed range of standard curve.

TABLE 7. Comparison of Hagahaiplasma and plasma-equivalent blood spot results:Means and standard errors (SE) for plasma (Pl)

and plasma-equivalent blood spot (P-E BS) values,Pearson correlation coefficients (r), r-squares (r2),

and P values from simple linear regressions,and paired t-test P values

Mean (SE) r r2 P t-test

A (ng/mL)Pl 0.60 (0.41)P-E BS 0.65 (0.36) 0.90 0.81 ,0.01 0.11

LH (IU/L)Pl 4.55 (11.8)P-E BS 4.82 (13.7) 0.99 0.98 ,0.01 0.48

T (ng/dL)Pl 194.5 (146.7)P-E BS 214.8 (167.8) 0.98 0.95 ,0.01 0.05


a basis for cross-sex and cross-populationcomparison of pubertal timing and progres-sion. Accordingly, one of our first studiesemploying blood spots concerned character-ization of endocrine and morphologic statusover the juvenile period. The study, under-taken in collaboration with Nicholas Blur-ton Jones and with field collections by Jenni-fer Phillips-Davids (Phillips et al., 1991),involved the Hadza, a hunter-gatherer popu-lation of around 600 living in northwestTanzania and subject of intensive ongoinginvestigation by Nicholas Blurton Jones,Kristin Hawkes, and colleagues (BlurtonJones et al., 1989; Blurton Jones, 1993). TheHadza are of interest in part because, unlikethe intensively care-giving !Kung, they pur-sue a parental care strategy predicated onearly child foraging and substantial self-provisioning by the end of the first decade(Blurton Jones, 1993). Like the !Kung, sexpreference in child treatment is reported asminimal. Further, as government policy in-creasingly constrains Hadza to abandon for-aging for farming, collection of baselinepresedentization data was essential to pro-vide a comparative basis for future studiesof effects of changes in subsistence andlifestyle. Ease of collection, storage, andtransport of blood spots allows such rapidcross-sectional surveys.

In the last 2 weeks of the 1991 summerfield session, blood spot samples were col-lected from all available individuals ages5–20 (n 5 156). Samples were stored at am-bient temperatures but away from heatsources for up to 24 days before return to ourlaboratory for storage at 223°C until analy-sis. Unsurprising in the light of reports onsimilar groups in the region was our findingthat the Hadza are later maturing, showinga median age at significant elevation of LH(and inferentially of pubertal onset) ofaround 11.5 years in girls. But our finding ofa dramatic delay of pubertal onset in boys(median age at significant LH elevationabout 13.5 years), at least 2 years after girls(Fig. 1), was surprising and represents thelargest sex difference in pubertal onset thatwe have as yet observed. For instance, weand others find that, among American youth,boys reach endocrine and morphologic pu-berty about 6 months after girls (reviewed in

Angold and Worthman, 1993). Reports onsecular trends to accelerated reproductivematuration are based largely on age atmenarche (Eveleth and Tanner, 1990). Bloodspot sampling offers a feasible method totrack the developmental process and varia-tion in both sexes and to probe the social,cultural, and ecological factors and changesthat underlie such variation.

Longitudinal, population-basedepidemiologic study

Use of blood spot sampling has allowed usto undertake the first large-scale population-based endocrine study of the relationship ofnormal puberty to emergence of sex differ-ences and adult patterns of psychopathologyand of the effect of environmental quality onnormal variation in pubertal developmentin Western adolescents (Angold and Worth-man, 1993; Worthman, 1995). The mainstudy was initiated by Jane Costello for aninvestigation of mental health service needsfor rural youth (Costello et al., 1996) andcomprises a probability-based sample of1,100 children, ages 9, 11, and 13 years,drawn from 11 counties of western NorthCarolina. Heights, self-rated Tanner scores,and blood spot hormonal measures of LH,FSH, T, E2, DHEA-S, and A are used toevaluate development. Children are inter-

Fig. 1. Plasma equivalent blood spot LH in Hadzafemales (--p--) and males (—X—). Points representage-grouped means (6SD, n 5 5–13).


viewed, sampled, and measured annually athome in conjunction with parallel parentalinterviews; blood spot sampling offered theonly feasible method for hormone samplingby field teams, by offering ease of collectionand transport, minimal instrumentation,and high child and parent acceptability.These features allowed incorporation of en-docrine measures in an interview-based pro-tocol concerned largely with ascertainingpsychopathologic symptomatology, familyfunctioning and socioeconomics, and mentalhealth service use. For each annual hor-monal profile, two finger pricks are per-formed 20 min apart to be averaged in lateranalysis for optimal representation of pulsa-tile hormones (Bain et al., 1988; Goldzieheret al., 1976) or analyzed separately for gaug-ing dynamic responses. Samples are dried atroom temperature and stored at 4°C for upto 2 weeks and then posted in express weeklyshipments to the laboratory for storage at223°C and subsequent assay.

We are currently in the third year of thisongoing study, but our blood spot sampleanalyses have so far provided the basis forcharacterizing the relationship of puberty to

emerging sex difference in rates of depres-sion (Angold et al., 1995), for probing theconsiderable degree of individual physi-ologic variation and its correlates (McDadeet al., 1995; Stallings et al., 1995), and evenfor examining cortisol reactivity to noveltyand physical stress in relation to immunefunction and rearing conditions (McDade etal., 1997). These epidemiologic data from awell-fed, healthy American population ofadolescents also provide a substantial em-pirical basis for population comparisons. Byillustration, in Figure 2, the endocrine pro-file of Hadza youth are overlaid with meansand ranges for our North Carolina sample.Comparison of blood spot plasma equivalentvalues, in this case for LH, shows thatHadza of both sexes enter puberty well afterAmerican children. Increases in mean LHoccur later among Hadza, by about 2 yearsin girls and nearly 3.5 years in boys; thecontrast of Hadza and American boys under-scores the markedly delayed pubertal onsetof the former. We do not have a readyexplanation for pronounced male develop-mental delay among Hadza, though aspectsof behavior, social ecology, or sex-differenti-

Fig. 2. Plasma equivalent blood spot LH for Hadza youth (---X---) compared to a large sample inwestern North Carolina (GSMS, Q). Note that the points for Hadza represent age-grouped means (6SD,n 5 5–13) by sex from cross-sectional data and have been connected to aid readability. Points forAmericans represent ungrouped age-specific means (6SD) from a single year (year 1) of data collection.


ated sensitivity to specific conditions mayplay a role. Nonetheless, without the endo-crine data, we would not have distinguishedthis phenomenon so readily, nor would wehave been able to distinguish whether it wascentrally mediated (based on LH levels) ordue to gonadal refractoriness (based on go-nadal output); ability to monitor both pitu-itary regulatory hormones (LH in this case)as well as those of peripheral target organs(such as T or E2) allows discriminationacross levels of endocrine regulation. Bloodspot sampling provides a window to the setpoints and dynamics of brain regulation ofreproductive maturation and adult function.

Dynamic study: Serial measuresof endocrine response

Reproductive ecology is concerned withbehavior, both in terms of its direct impacton reproductive function (Cumming et al.,1994) and vice versa (Campbell and Leslie,1995). Despite the centrality of behavior as abridge between culture or human ecologyand reproductive performance or fertility,current reproductive ecology lacks endo-crine studies of behavior-biology interac-tions. Blood spot measures facilitate suchstudies because they allow serial samplingto monitor central and peripheral dynamicsboth acutely and through time.

For reproductive ecology, a classic in-stance of biobehavioral dynamics is the sup-pressive effect of breast-feeding on ovarianfunction, a centrally mediated effect re-flected in patterns of suckling-induced PRLrelease. In collaboration with CatherinePanter-Brick, we recently used blood spotsampling to examine the postsuckling trajec-tory of circulating PRL in two sympatricNepali groups, the Tamang and Kami. Moth-ers in these two groups exhibit similarbreast-feeding frequency and timing ofsupplementation but have quite differentinterbirth intervals: Tamang women, de-spite their greater height and weight, haveaverage birth intervals 8 months longerthan Kami (29 months) (Panter-Brick, 1991).The disparity may be due in part to heavyworkloads and cyclic weight loss amongTamang women (Panter-Brick et al., 1993),but the group difference in duration of birthspacing (Panter-Brick, 1991) suggested a

role for differences in endocrine effects ofnursing.

In August 1991, we obtained serial finger-prick blood spot samples from all availablenursing mothers (55 Tamang and 17 Kami)at 5, 30, and 50 min after termination of anursing bout. Samples were air-dried andstored with drierite at ambient temperatureuntil shipment to the laboratory, where theywere stored at 223°C within 3 weeks aftercollection. Predictably, the magnitude andduration of suckling-induced PRL increasedecreased with time postpartum, but groupsdiffered in the rate of both declines (Fig. 3).Pituitary responsiveness to suckling de-cayed more rapidly among Kami both withtime after nursing and by infant age. Under-lying group differences in neuroendocrinefunction are contraindicated by the absenceof group difference in immediate postsuck-ling (5 min) prolactin levels among motherswith young infants (#11 mos.), though Kamishowed significantly lower PRL by 50 minpostsuckling in mothers with infants lessthan 22 months old. Further analyses sug-gested that PRL levels relate to differentvariables for the two groups, extrinsic (nurs-

Fig. 3. Circulating prolactin concentrations(mean 6 s.e.) at specified intervals after suckling (5, 30,and 50 min) by infant age (#11, 11–#15, and .15months) in two sympatric Nepali groups, the Tamang(---N---) and the Kami (—X—).


ing bout length, infant age) for Tamang andintrinsic (maternal age) for Kami (Stallingset al., in press). We found that postsucklingPRL levels were highly predictive of mainte-nance of postpartum amenorrhea: Tamangand Kami women who maintained elevatedPRL levels ($10 ng/ml) as long as 50 minpostsuckling were five times more likely tobe amenorrheic (Stallings et al., 1996).

Serial hormone measures offer consider-able potential for characterizing the role ofbehavior in reproductive ecology in men(Worthman and Konner, 1987) as well aswomen. We have recently completed collec-tions for a study with Virginia Vitzthum ofacute endocrine responses to exercise bymenstrual cycle phase among Bolivianwomen at high altitude. This experience andthat with Cynthia Beall in collections alsoundertaken in Bolivia (Worthman et al.,1997) suggest that finger-prick sampling isacceptable even to people in Andean regionslong known for aversion to venipuncture.

Comparative life span endocrinology

Comparative endocrine analyses informgrowing interest in the underlying social,cultural, and ecological factors that influ-ence variation in the ontogeny of reproduc-tive function. Recent studies of nonpreg-nant, nonlactating women across populationshave revealed significant variation in aver-age levels of ovarian hormones (Ellison,1994), and intrapopulation studies havelinked ovarian hormone variation to differ-ences in energy balance (weight, workload)(Ellison, 1994; Ellison et al., 1989; Jasien-ska and Ellison, 1993; Panter-Brick et al.,1993). Comparative studies have shown thatdifferences in the timing and course of puber-tal development can be linked to variation inenvironmental quality and that such ecologi-cal variation can be quite marked acrosspopulations (Worthman, 1995), within popu-lations (Riley, 1994; Laska-Mierzejewska,1995; Stallings et al., 1995; Wood, 1994), andeven within families or households (Worth-man, 1996, in press).

Cumulatively, these studies suggest thevalue to reproductive ecology of a life his-tory, life span approach which incorporatesreproductive ontogeny with adult functionand aging. Such a perspective addresses

three related issues in relation to theirimpact on central neuroendocrine regula-tion and gonadal function: 1) how ecologicaland behavioral factors affect the onset andtrajectory of reproductive development, 2)how divergent developmental histories con-dition the effect of ecological and behavioralfactors on adult reproductive function, and3) how these histories shape processes ofsenescence. This contextual ontogenetic ap-proach should help determine whether diver-gent life histories condition the effect ofproximal factors on reproductive processesthat generate variation in fertility.

Blood spot sampling, for both its conve-nience and range of readable hormones,enhances the ability to study groups fromwidely variant environments and reproduc-tive life histories and thus promotes com-parative endocrinological research. The po-tential of this approach can be exemplifiedby a comparison of two populations of PapuaNew Guinea, the Hagahai and the Amele,studied in collaboration with Carol Jenkinsof the IMR (Jenkins et al., 1989; Worthmanet al., 1993). The Hagahai, described brieflyabove, are a demographically declining popu-lation plagued by steep morbidity and mor-tality rates and poor nutrition, particularlyamong women and children; these condi-tions are reflected in stunted growth, latematuration, and small adult body size. TheAmele, by contrast, are a numerous sub-coastal people who enjoy relatively goodhealth care and nutrition, lower mortality,and concomitantly earlier maturation.Through ongoing work by Jenkins and theIMR, we had assembled a large series ofplasma samples from Hagahai, and we usedblood spot sampling to rapidly obtain across-sectional age-stratified sample set forAmele. Sample collection was initiated byJim Rilling of our laboratory and the surveycollected by Daina Lai of the IMR in August1994. Such cross-sectional survey approachesare useful for studying juveniles and men,but the variability of endocrine profile inrelation to women’s reproductive state prob-lematizes a survey approach to characteriza-tion of reproductive endocrinology of womenin the reproductive years. Reproductive hor-mone values for young Hagahai (n 5 21)and Amele (n 5 36) men ages 21–25 years


are plotted in Figure 4. Selection of this agerange avoids the issue of population differ-ences in maturational schedule and allowscomparison of the hypothalamo-pituitary-gonadal (HPG) axis at the point of peakyoung adult function. The population differ-ence is readily apparent: young Amele menhave twice the testosterone and the LH oftheir Hagahai counterparts. That the popu-lation difference in T and LH is symmetricalis significant, for it indicates that the dispar-ity in testicular output is centrally medi-ated, not a product of gonadal refractorinessor impairment. In other words, the HPGaxis is downregulated among Hagahai rela-tive to Amele men.

These observations furthermore demon-strate population differences in gonadal out-put of steroids known to be linked not only tobehavior and cognition but also to long-termhealth outcomes, including risk for reproduc-tive cancers. Hagahai men clearly will haveless quantitative lifetime exposure to testos-terone and its metabolites than will Amele,not only because they mature later than doAmele but also because testicular output isregulated to a lower level. Similar effectsmay obtain in women. Dramatically escalat-ing rates of reproductive cancers, particu-larly breast cancer (Kelsey and Horn-Ross,1993), have spurred concern with the social-ecological transformations that may impairwomen’s reproductive function (Lasley et

al., 1994) and heighten risk for neoplasia(Eaton et al., 1994). Population variation inbreast cancer rates has been related todifferences in serum levels of ovarian hor-mones (Pike et al., 1993; Bernstein andRoss, 1993). Comparisons of premenopausalWestern women and Asian women haveoverwhelmingly found higher estrogen lev-els in Western women (Bernstein and Ross,1993). Recent anthropological studies of Lesewomen of Zaire, the Tamang of Nepal, andrural farm women in southern Poland havefound lower average mid-luteal progester-one levels than among Boston women ofsimilar ages (Ellison, 1994). Differences indiet or energy expenditure may account forsome of these divergent endocrine patterns(Ellison, 1994; Bernstein and Ross, 1993).

Recognition of population differences andwithin-population secular trends in gonadalsteroid exposure has raised keen interest inthe magnitude, causes, and consequences ofsuch variation. The blood spot SHBG methodreported here should also assist such investi-gation, as levels of this binding proteinstrongly influence bioavailable circulatingsteroid concentrations. Indeed, variation inSHBG values has been linked to breastcancer risk. Lower levels of SHBG have beenfound in high-risk groups, such as obesewomen (Anderson, 1974; Bernstein and Ross,1993; Moore et al., 1987), women with earlyage at menarche and established regularityof menstrual cycles (Apter and Vihko, 1989;Bernstein and Ross, 1993; Siiteri and Sim-berg, 1986), and nulliparous vs. parous pre-menopausal women (Bernstein et al., 1985;Moore et al., 1987). Although some associa-tions have been found between breast can-cer and progesterone, PRL, and the andro-gens T and DHEA-S, studies of thesehormones are less extensive and findingsless consistent (Bernstein and Ross, 1993).

Beyond reproductive ecology

Applications of the blood spot samplingmethods described here, available else-where, or currently under development inour laboratory contribute to biosocial andbehavioral research outside reproductiveecology as well. To briefly indicate the scopeof such applications, we highlight four areasof investigation in stress, immune function,

Fig. 4. Group variation in hypothalamo-pituitary-gonadal set points, showing differences between twoPapua New Guinea groups, Hagahai and Amele, in T(---W---) and LH (—Q—) concentrations (mean 6 SD) inyoung men ages 21–25 years.


metabolism, and multiple or cross-axis endo-crine function. First, concerning stress, theblood spot cortisol measure reported herecan be deployed for population comparisonsor dynamic studies of individuals. For in-stance, the 20 min dual blood spot samplingprotocol used in our North Carolina study,described above, also allows characteriza-tion of adrenocortical responses to experi-ence. Because we sample at 20–30 min intothe home interview and a cortisol responsetakes about 15–20 min to peak in circula-tion, our sample at time one provides ameasure of response to social novelty (viz.,unknown interviewers coming into thehome), while that at time two indexes theresponse to the stress of the finger prick attime one. We have found strong sex differ-ences and developmental effects in degree ofcortisol responsiveness to social novelty (Mc-Dade et al., 1997). Thus, careful samplingdesign that takes advantage of the interven-tions introduced by the research protocolsthemselves can provide a window to thedynamics of stress responsiveness.

Although saliva sampling offers real ad-vantages for stress research and is com-monly used in this area (Pollard, 1995),blood spot sampling allows a wider array ofphysiologic end points to be probed in tan-dem. For example, large proteins such asantibodies are also stable and measurable inblood spots, which opens a window to im-mune function, a central actor in adaptationto pathogens and thus a core determinant ofwell-being and disease vulnerability. Mea-sures of cortisol in the North Carolina studywere coupled with those of antibody to resi-dent Epstein Barr virus that varies in in-verse proportion with immune competenceto reveal an effect of environmental qualityon immunocompetence (McDade et al., 1997).Psychosocial and material adversity wasfound to depress immunocompetence, butboth strong sex differences and developmen-tal effects on this relationship were alsoidentified. Another published use of proteinmeasures in blood spots for anthropologicalpurposes is for ABO blood group typing(Aebischer et al., 1990).

Third, more could be done with the well-established measures of thyroid function(TSH, T3, T4) to track metabolic regulation

as a regulator of energy consumption, amarker of nutritional deficiency, and a factorin aging. As a simple instance of this applica-tion, our measure of blood spot TSH forCynthia Beall’s study of adaptation to highseasonal variation in food intake and ther-mal loads in Tibetans allowed us to establishthat hypothyroidism could not explain thelack of seasonal variation in basal metabolicrates in this population (Beall et al., 1996).

Lastly, the wide range of hormones mea-surable in blood spots facilitates expansionof research on endocrine regulation, adapta-tion, and response to comparative analysisacross endocrine axes. Reproductive ecologyhas scarcely posed the question of whetheror how effects of nutritional constraints ongonadal function may be moderated or evenmediated by the ability to make metabolicadjustments. Nor have we consideredwhether individual, intrapopulation varia-tion in reproductive hormone function orreproductive output may correlate with theadrenal androgen markers (DHEA, DHEAS)linked to longevity. Outside of reproductiveecology, the literature on stress documents arelationship between temperamental differ-ences in cortisol responsiveness to differ-ences in neuroendocrine functioning (sero-toninergic and noradrenergic) that in turnrelate to differential morbidity and mortal-ity risks that are both behaviorally andphysiologically mediated (Barr et al., 1994;Williams, 1994). Comparative studies ofthese dynamics, and of the effects of varyingsocial and ecological environments on neuro-endocrine function, temperament, and re-sulting differential well-being may be morefeasible with blood spot sampling. A primaryfunction of endocrine systems is to set physi-ologic priorities and allocate limited re-sources within the body to meet currentfunctional demands and constraints; thecommon occurrance of behavioral, psychoso-cial, or other environmental challengesmeans that demands of all physiologic sys-tems can rarely be met simultaneously butmust be juggled, traded off, and optimizedover the short, medium, and long term.Investigation of genetic and facultative ordevelopmentally induced variation in rela-tionships among regulatory axes with re-spect to differential performance or well


being is still in its infancy (but see Finch andRose, 1995) but bids fair to yield a morecomplete explanatory framework for endo-crine function.

CONCLUSIONS

The ability to measure hormones acrosspopulations living in diverse physical andsocial environments presents an opportu-nity to probe the extent, causes, and conse-quences of human variation in endocrinephysiology and to expand our understand-ing of ‘‘normal’’ human endocrine function ineveryday life. The present report documentsmethodological details, sample stabilities,assay performance, and other technical as-pects of the sampling and assay methods.We have illustrated the research advan-tages of finger-prick sampling and bloodspot assays for diverse study designs, includ-ing rapid cross-sectional surveys, large epi-demiologic and longitudinal studies, serialsampling, and self-sampling by subjects.The availability of highly sensitive, accu-rate, and precise assays for determination ofa wide range of hormones (gonadotropins,gonadal and adrenal steroids, and SHBG)from a very small blood sample (200 µl ofwhole blood) provides the opportunity forminimally invasive, maximally convenientascertainment of developmental or reproduc-tive status and an index of bioavailablegonadal steroids.

Population variation represents an impor-tant source of information about biologicalmechanisms, about responsiveness of theendocrine system to environmental influ-ences, about ontogenetic and acute sourcesof variation, and about the consequences ofthis variation for human function and well-being. Furthermore, blood spot sampling,especially self-collections, provides an oppor-tunity to integrate hormonal studies into awide range of designs, from small samplesfrom remote or low-tech anthropological fieldsettings to public health surveillance pro-grams that provide large data bases ondemographics, nutrition, socioeconomics, andreproductive histories. Thus, availability ofpractical collection methods and improvedhormonal assays permit identification of en-vironmental, social, occupational, and behav-

ioral effects on reproductive processes andreproductive health in particular, while theyalso expand the potential scope of endocrineresearch in general.

ACKNOWLEDGMENTS

We gratefully acknowledge the collabora-tion of several investigators central to thefield studies described here. They are AdrianAngold, Cynthia Beall, Nicholas BlurtonJones, Ben Campbell, Jane Costello, DavidGubernick, Carol Jenkins, and CatherinePanter-Brick. We also thank Patricia Castro,Daina Lai, Jennifer Phillips Davids, andJim Rilling for excellent technical or fieldassistance. Finally, we recognize the essen-tial contributions of the people who partici-pated in the studies: the Amele, Hagahai,Hadza, Tamang and Kami, and Americansof western North Carolina and of centralWisconsin. This work was partially sup-ported by a W.T. Grant Foundation facultyscholarship to C.M.W.; W.T. Grant Founda-tion grants 92-1489-92 and 94-1489-2; Uni-versity Research Committee of Emory Uni-versity; and NIMH MH48085 (to Costello).Reagents for initial development of the PRL,FSH, and LH assays were donated by Wal-lac, Inc.

LITERATURE CITED

Aebischer ML, Martorana MC, Costa F, Battaggia C,Madera A, Destito D, Machera F, Bailly C, andAngeloni P (1990) Evaluation of the sensitivity ofmicrofilter paper assays in an anthropological study:Results of samples from Cameroon and Tanzania.Anthropol. Anz. 48:15–23.

Anderson D (1974) Sex-hormone-binding-globulin. Clin.Endocrinol. 3:69–96.

Angold A, and Worthman CM (1993) Puberty onset ofgender differences in rates of depression: A develop-mental, epidemiologic and neuroendocrine perspec-tive. J. Affect. Disord. 29:145–158.

Angold A, Costello EJ, and Worthman CM (1995) Sex-differentiated effects of puberty on the relationshipbetween adversity and symptoms of depression: Areport from the Great Smoky Mountains Study. Am. J.Phys. Anthropol. 20(Suppl.):58 (abstract).

Apter D, and Vihko RJ (1989) Some endocrine character-istics of early menarch, a risk factor for breast cancer,are preserved into adulthood. Int. J. Cancer 44:783–787.

Bain J, Sanders RM, Langevin R, Hucker S, and D’CostaM (1988) Serum pituitary and steroid hormone levelsin the adult male: One value is as good as the mean ofthree. Fertil. Steril. 49:123–126.

Barr RG, Boyce WT, and Zeltzer LK (1994) The stress-illness association in children: A perspective from the


biobehavioral interface. In RJ Haggerty, LR Sherrod,M Garmezy, and M Rutter (eds.): Stress, Risk, andResilience in Children and Adolescents. Cambridge:Cambridge University Press, pp. 182–224.

Bassett F, Gross B, and Eastman C (1986) Radioimmu-noassay of prolactin in blood spotted on filter paper.Clin. Chem. 32:854–856.

Beall C, Worthman CM, Stallings J, Strohl K, Britten-ham G, and Barragan M (1992) Salivary testosteroneconcentration of Aymara men native to 3600 m. Ann.Hum. Biol. 19:67–78.

Beall CM, Henry J, Worthman CM, and Goldstein MC(1996) Basal metabolic rate and dietary seasonalityamong Tibetan nomads. Am. J. Hum. Biol. 8:361–370.

Bernstein L, and Ross R (1993) Endogenous hormonesand breast cancer risk. Epidemiol. Rev. 15:48–65.

Bernstein L, Pike MC, Ross RK, Judd HL, Brown JB,and Henderson BE (1985) Estrogen and sex hormone–binding globulin levels in nulliparous and parouswomen. J. Natl. Cancer Inst. 74:741–745.

Blank B, Attanasio A, Rager K, and Gupta D (1978)Determination of serum sex hormone binding globulin(SHBG) in preadolescent and adolescent boys. J.Steroid Biochem. 9:121–125.

Blurton Jones N (1993) The lives of hunter-gathererchildren: Effects of parental behavior and parentalreproductive strategy. In ME Pereira and LA Fair-banks (eds.): Juvenile Primates. New York: OxfordUniversity Press, pp. 309–326.

Blurton Jones N, Hawkes K, and O’Connell JF (1989)Measuring and modelling costs of children in twoforaging societies: Implications for schedule of repro-duction. In V Standen and R Foley (eds.): Compara-tive Socioecology of Mammals and Humans. Oxford:Blackwell, pp. 367–390.

Brombacher PJ, Henkel E, Dessler AC, and Van Dieijen-Visser MP (1988) Multicentre study of a new test forTSH-screening in blood spots. Ann. Clin. Biochem.25:530–535.

Campbell BC, and Leslie PW (1995) Reproductive ecol-ogy of human males. Yearb. Phys. Anthropol. 21:1–26.

Campbell KL (1994) Blood, urine, saliva, and dip-sticks:Experiences in Africa, New Guinea, and Boston. Ann.N. Y. Acad. Sci. 709:312–330.

Carter G, Holland S, Alaghband-Zadeh J, Rayman G,Dorrington-Ward P, and Wise P (1983) Investigationof hirsutism: Testosterone is not enough. Ann. Clin.Biochem. 20:262–263.

Costello EJ, Angold A, Burns B, Stangl DK, Tweed DL,Erkanli A, and Worthman CM (1996) The GreatSmoky Study of Youth: Goals, design, methods, andthe prevalence of DSM-III-R disorders. Arch. Gen.Psychiatry 53:1129–1136.

Cumming D, Wheeler G, and Harber V (1994) Physicalactivity, nutrition, and reproduction. Ann. N. Y. Acad.Sci. 709:55–76.

Dunkel L, Alfthan H, Stenman UH, Tapanainen P, andPerheentupa J (1990) Pulsatile secretion of LH andFSH in prepubertal and early pubertal boys revealedby ultrasensitive time-resolved immunofluorometricassays. Pediatr. Res. 27:215–219.

Dunn J, Nisula B, and Rodbard D (1981) Transport ofsteroid hormones. Binding of 21 endogenous steroidsto both testosterone-binding globulin and corticoste-roid-binding globulin in human plasma. J. Clin. Endo-crinol. Metab. 53:58–68.

Eaton S, Pike M, Short R, Lee N, Trussell J, Hatcher R,Wood J, Worthman C, Blurton Jones N, Konner M,Hill K, Bailey R, and Hurtado A (1994) Women’sreproductive cancers in evolutionary context. Q. Rev.Biol. 69:354–367.

Egan S, Betts P, Thomson S, Wallace AM, and Wood PJ(1989) A blood spsot androstenedione assay suitablefor home monitoring of steroid replacement therapy incongenital adrenal hyperplasia. Ann. Clin. Biochem.26:262–267.

Ellison P (1988) Human salivary steroids: Methodologi-cal considerations and applications in physical anthro-pology. Yearb. Phys. Anthropol. 31:115–142.

Ellison P (1994) Salivary steroids and natural variationin human ovarian function. Ann. N. Y. Acad. Sci.709:287–298.

Ellison P, Peacock N, and Lager C (1989) Ecology andovarian function among Lese women of the IturiForest, Zaire. Am. J. Phys. Anthropol. 78:519–526.

Eveleth PB, and Tanner JM (1990) Worldwide Variationin Human Growth, 2nd ed. Cambridge: CambridgeUniversity Press.

Finch CE, and Rose MR (1995) Hormones and thephysiological architecture of life history evolution. Q.Rev. Biol. 70:1–52.

Goldzieher JW, Tazewell SD, Smith KD, and Stein-berger E (1976) Improving the diagnostic reliability ofrapidly fluctuating plasma hormone levels by opti-mized multiple-sampling techniques. J. Clin. Endocri-nol. 43:823–830.

Haavisto AM, Dunkel L, Pettersson K, and HuhtaniemiI (1990) LH measurements by in vitro bioassay anda highly sensitive immunofluorometric assay improvethe distinction between boys with constitutional de-lay of puberty and hypogonadism. Pediatr. Res. 27:211–214.

Hofman LF, Klaniecki JE, and Smith EK (1985) Directsolid-phase radioimmunoassay for screening 17 alpha-hydroxyprogesterone in whole-blood samples fromnewborns. Clin. Chem. 31:1127–1130.

Jasienska G, and Ellison P (1993) Heavy workloadimpairs ovarian function in Polish peasant women.Am. J. Phys. Anthropol. 16:117–118 (abstract).

Jenkins C, Dimitrakakis M, Cook I, Sanders R, andStallman N (1989) Culture change and epidemiologi-cal patterns among the Hagahai, Papua New Guinea.Hum. Ecol. 17:27–57.

Kelsey J, and Horn-Ross P (1993) Breast cancer: Magni-tude of the problem and descriptive epidemiology.Epidemiol. Rev. 15:7–16.

Knudsen RC, Slazyk WE, Richmond JY, and HannonWH (1993) Guidelines for the shipment of dried bloodspot specimens. Infant Screening 16:1–4.

Kraiem Z, Kahana L, Elias V, Ghersin S, and SheinfeldM (1980) Radioimmunoassay of cortisol in blood col-lected on filter paper. Clin. Chem. 26:1916–1917.

Laska-Mierzejewska T (1995) Age at menarche as anindicator of the socioeconomic situation of rural girlsin Poland in 1967, 1977, and 1987. Am. J. Hum. Biol.7:651–656.

Lasley B, Mobed K, and Gold E (1994) The use of urinaryhormonal assessments in human studies. Ann. N. Y.Acad. Sci. 709:299–311.

Lovgren T, Hemmila I, Pettersson K, and Halonen P(1985) Time-resolved fluorometry in immunoassay. InWP Collins (ed.): Alternative Immunoassays.Chichester: John Wiley and Sons, pp. 203–217.

Malamud D, and Tabak L (eds.) (1993) Saliva as aDiagnostic Fluid. New York: New York Academy ofSciences.

Marshall J, Dalkin A, Haisenleder D, Paul S, OrtolanoG, and Kelch R (1991) Gonadotropin-releasing hor-mone pulses: Regulators of gonadotropin synthesisand ovulatory cycles. Rec. Prog. Horm. Res. 47:155–198.

McDade T, Worthman CM, Angold A, Costello EJ, andStallings, J (1995) Physiologic bases of individual


variation in pubertal timing and progression: Reportfrom the Great Smoky Mountains Study. Am. J. Phys.Anthropol. 20(Suppl.):148 (abstract).

McDade TW, Stallings JF, and Worthman CW (1997)Psychosocial stress and cell-mediated immune func-tion: validation of a blood spot method for Epstein-Barr virus antibodies. Am. J. Phys. Anthropol. Suppl.24:164–5.

Meredith NK, and Hannon WH (1993) Preparation ofdried blood spot materials for quality assurance ofassays for antibodies to human immunodeficiencyvirus. In BL Therrell (ed.): Laboratory Methods forNeonatal Screening. Washington, DC: American Pub-lic Health Association, pp. 225–241.

Moore JW, Key TJA, Bulbrook RD, Clark GM, Allen DS,Wang DY, and Pike MC (1987) Sex hormone bindingglobulin and risk factors for breast cancer in a popula-tion of normal women who had never used exogenoussex hormones. Br. J. Cancer 56:661–666.

National Committee for Clinical Laboratory Standards(NCCLS) (1992) NCCLS Approved Standard LA4-A2.Blood Collection on Filter Paper for Neonatal Screen-ing Programs. Illanova, PA: National Committee forLaboratory Standards.

Osredkar J, Vrhovec I, Jesenovec N, Kocijancic A, andPrezelj J (1989) Salivary free testosterone in hirsut-ism. Ann. Clin. Biochem. 26:522–526.

Panter-Brick C (1991) Lactation, birth spacing andmaternal work-loads among two castes in rural Nepal.J. Biosoc. Sci. 23:137–154.

Panter-Brick C, Lotstein DS, and Ellison PT (1993)Seasonality of reproductive function and weight lossamong rural Nepali women. Hum. Reprod. 8:684–690.

Pardridge W (1981) Transport of protein-bound hor-mones into tissues in vivo. Endocr. Rev. 2:102–123.

Petra P (1991) The plasma sex steroid binding protein(SBP or SHBG). A critical review of recent develop-ments on the structure, molecular biology and func-tion. J. Steroid Biochem. Mol. Biol. 40:735–753.

Petsos P, Ratcliffe WA, Heath DF, and Anderson DC(1986) Comparison of blood spot, salivary and serumprogesterone assays in the normal menstrual cycle.Clin. Endocrinol. 24:31–38.

Phillips J, Worthman CM, and Stallings JF (1991) Newfield techniques for detection of female reproductivestatus. Am. J. Phys. Anthropol. 85:143 (abstract).

Pike M, Spicer D, Dahmoush L, and Press M (1993)Estrogens, progestogens, normal breast cell prolifera-tion, and breast cancer risk. Epidemiol. Rev. 15:17–35.

Pollard T (1995) Use of cortisol as a stress marker:Practical and theoretical problems. Am. J. Hum. Biol.7:265–274.

Riad-Fahmy D, Read G, Walker R, Walker S, andGriffiths K (1987) Determination of ovarian steroidhormone levels in saliva: An overview. J. Reprod. Med.32:254–272.

Riley A (1994) Determinants of adolescent fertility andits consequences for maternal health, with specialreference to rural Bangladesh. Ann. N. Y. Acad. Sci.709:86–100.

Rilling JK, Worthman CM, Campbell BC, Stallings JF,and Mbizva M (1996) Ratios of plasma and salivarytestosterone throughout puberty: production vs. bio-availability. Steroids 61:374–378.

Ruutiainen K, Sannikka E, Sanntti R, Erkkola R, andAdlercreutz H (1987) Salivary testosterone in hirsut-ism: Correlations with serum testosterone and thedegree of hair growth. J. Clin. Endocrinol. Metab.65:1015–1020.

Sannikka E, Terho P, Suominen J, and Santti R (1983)Testosterone concentrations in human seminal plasma

and saliva and its correlation with non-protein boundand total testosterone levels in serum. Int. J. Androl.6:319–330.

Siiteri P, Murai J, Hammond G, Nisker J, Raymoure W,and Kuhn R (1982) The serum transport of steroidhormones. Rec. Prog. Horm. Res. 38:457–503.

Siiteri PK, and Simberg NH (1986) Changing conceptsof active androgens in blood. Clin. Endocrinol. Metab.15:247–258.

Stallings J, Angold A, Costello EJ, and Worthman CM(1995) Environmental effects on pubertal develop-ment in western adolescents: Report from the NorthCarolina Great Smoky Mountain Study. Am. J. Phys.Anthropol. 20(Suppl.):201 (abstract).

Stallings JF, Worthman CM, Panter-Brick C, and CoatesRJ (1996) Prolactin response to suckling and mainte-nance of postpartum amenorrhea among intensivelybreastfeeding Nepali women. Endocr. Res. 22:1–28.

Stallings JF, Worthman CM, and Panter-Brick C (inpress) Biological and behavioral factors influence castedifferences in prolactin levels among breastfeedingNepali women. Am. J. Hum. Biol.

Stenman UH, Alfthan H, Koskimies A, Seppala M,Pettersson K, and Lovgren T (1985) Monitoring theLH surge by ultrarapid and highly sensitive immuno-fluorometric assay. Ann. N. Y. Acad. Sci. 442:554–550.

Thomson S, Wallace AM, and Cook BA (1989) 125I-radioimmunoassay for measuring androstenedione inserum and in blood-spot samples from neonates. Clin.Chem. 35:1706–1712.

Torresani TE, and Scherz T (1986) Neonatal thyroidscreening by a non-radioactive method: Evaluation ofthyrotropin time-resolved fluoroimmunoassay. Clin.Chem. 32:1013–1016.

Vining R, and McGinley R (1987) The measurement ofhormones in saliva: Possibilities and pitfalls. J. Ste-roid Biochem. 27:81–94.

Waite KV, Maberly GF, and Eastman CJ (1987) Storageconditions and stability of thyrotropin and thyroidhormones on filter paper. Clin. Chem. 33:853–855.

Wang C, Plymate S, Nieschlag E, and Alvin-Paulsen C(1981) Salivary testosterone in men: Further evidenceof a direct correlation with serum testosterone. J.Clin. Endocrinol. Metab. 53:1021–1024.

Williams RB (1994) Neurobiology, cellular and molecu-lar biology, and psychosomatic medicine. Psychosom.Med. 56:308–315.

Wood JW (1994) Dynamics of Human Reproduction:Biology, Biometry, Demography. Hawthorne, NY: Al-dine de Gruyter.

Worthman CM (1995) Epidemiology of human develop-ment. Am. J. Hum. Biol. 7:138–139 (abstract).

Worthman CM (1996) Biosocial determinants of sexratios: survivorship, selection, and socialization in theearly environment. In SJ Ulijaszek and CJK Henry(eds.): Long-Term Consequences of Early Environ-ments. Cambridge: Cambridge University Press, pp.45–68.

Worthman CM (in press) Evolutionary perspectives onthe onset of puberty. In WR Trevathan, JJ McKenna,and EO Smith (eds.): Evolutionary Medicine. OxfordUniversity Press.

Worthman CM, and Konner MJ (1987) Testosteronelevels change with subsistence hunting effort in !KungSan men. Psychoneuroendocrinology 12:449–458.

Worthman CM, and Stallings JF (1994) Measurement ofgonadotropins in dried blood spots. Clin. Chem. 403:448–453.

Worthman CM, Stallings JF, and Gubernick D (1991)


Measurement of hormones in blood spots: A non-isotopic assay for prolactin. Am. J. Phys. Anthropol.85:186–187.

Worthman CM, Jenkins C, Stallings JF, and Lai D(1993) Atenuation of nursing-related ovarian suppres-sion and high fertility in well-nourished, intensivelybreast-feeding Amele women of lowland Papua NewGuinea. J. Biosoc. Sci. 25:425–443.

Worthman CM, Beall CM, and Stallings JF (1997)Population variation in reproductive function of men.Am. J. Phys. Anthropol. 24(Suppl):246.

Wu F, Butler G, Kelnar C, Stirling H, and Huhtaniemi I(1991) Patterns of pulsatile luteinizing hormone andfollicle-stimulating hormone secretion in prepubertal(midchildhood) boys and girls and patients with idio-pathic hypogonadotropic hypogonadism (Kallman’ssyndrome): A study using an ultrasensitive time-resolved immunofluorometric assay. J. Clin. Endocri-nol. Metab. 72:1229–1237.

Yalow R, and Berson S (1971) Principles of Competi-tive Protein Binding Assays. Philadelphia: JB Lippin-cott Co.


hormone measures in finger-prick blood spot samples: new...

Documents