Clinica Chimica Acta 305 (2001) 35–40www.elsevier.com/ locate /clinchim

Screening for primary hyperparathyroidism (PHPT) in clinicpatients: differential diagnosis between PHPT and

malignancy-associated hypercalcemia by routine blood testsa a , b c d*Seung Jin Kim , Eiichi Shiba , Ikuhiro Maeda , Toshiaki Yoshioka , Nobuyuki Amino ,

aShinzaburo NoguchiaDepartments of Surgical Oncology, Osaka University Medical School, 2-2-E-10 Yamadaoka, Suita City, Osaka 565-0871, Japan

bDepartment of Central Laboratory for Investigation, Osaka University Medical School, 2-2-E-10 Yamadaoka, Suita City,Osaka 565-0871, Japan

cDepartment of Urology, Osaka University Medical School, 2-2-E-10 Yamadaoka, Suita City, Osaka 565-0871, JapandLaboratory Medicine, Osaka University Medical School, 2-2-E-10 Yamadaoka, Suita City, Osaka 565-0871, Japan

Received 30 May 2000; received in revised form 24 October 2000; accepted 2 November 2000

Abstract

Screening for primary hyperparathyroidism (PHPT) by measurement of the serum calcium concentration detects onepatient per 500–1000 individuals in Western countries, and one patient per 2500–5000 subjects in Japan. Among clinicpatients, however, the presence of many false-positive cases due to malignancy-associated hypercalcemia (MAH) reducesthe benefit of such screening. We evaluated a new method of screening for PHPT based on the results of routine blood testsusing the hospital information system (HIS) at our hospital. This new method could distinguish PHPT from MAH. Thisstudy included 25 179 blood samples in which the serum calcium (Ca), albumin (Alb), chloride (Cl) and inorganic phosphate(IP) concentrations had been measured between March, 1994 and February, 1995 at Osaka University Medical Hospital. TheHIS was programmed to pick blood samples that satisfied Formula 1 [Ca(mEq/ml).0.33Alb(g /dl)14.1] and Formula 2h[Cl(mEq/ml)284]3[103Alb215]4[IP(mg/dl)43.1].400j. Of data from 25 179 blood samples collected, those from 54patients satisfied both Formulae 1 and 2. The patients from which these samples were derived from were subject to furtheranalysis: medical records were studied and the intact-parathyroid hormone concentration was measured if necessary. Of these54 cases, 19 patients (35.2%) were subsequently diagnosed with PHPT, including two, who were newly diagnosed withPHPT by this screening procedure. Although 35 (64.8%) of 54 patients were false-positive, many of them were treated withblood purification therapies in the Department of Pediatrics or the Intensive Care Unit (ICU). On the other hand, there werefour false-positive cases (7.4%) caused by MAH. False-negative case in this study was only one patient (5%), whosediagnosis was normocalcemic PHPT. When omitting samples from pediatric patients and those in ICU, this screeningprocedure for PHPT has the advantage of being able to differentiate this diagnosis from MAH. 2001 Elsevier ScienceB.V. All rights reserved.

Keywords: Hyperparathyroidism; Malignancy-associated hypercalcemia; Screening; Routine blood tests

Abbreviations: PHPT, primary hyperparathyroidism; MAH, malignancy-associated hypercalcemia; PTH, parathyroid hormone; PTHrP,PTH-related protein; DIC, disseminated intravascular coagulation syndrome

*Corresponding author. Tel.: 181-6-6879-3772; fax: 181-6-6879-3779.E-mail address: eshiba@onsurg.med.osaka-u.ac.jp (E. Shiba).

0009-8981/01/$ – see front matter 2001 Elsevier Science B.V. All rights reserved.PI I : S0009-8981( 00 )00403-4

36 S.J. Kim et al. / Clinica Chimica Acta 305 (2001) 35 –40

1. Introduction was considered to have likely been obtained from aPHPT patient if the chloride, albumin and phosphate

Primary hyperparathyroidism (PHPT) is not as concentrations satisfied Formula 2 in Table 1, intro-rare as once believed. Since the introduction of duced by Lind and Ljunghall [13]. Their originalscreening using serum calcium, the incidence of formula is as follows: (Cl284)3(Alb215) / IP (ClPHPT in the general population has been reported to and IP; mmol / l, Alb; g / l). A borderline range (400–be 1:500–1000 in Europe and the US. Among 500) was indicated in their study and 96% of HPTpostmenopausal women, the incidence of PHPT has patients revealed more than 500 in the formula.been reported to be 2–3% [1–8]. Although the However, in order to reduce the number of falseprevalence of PHPT in Japan is lower than in Europe negatives in this study, the cutoff value was selectedand the US, with the routine measurement of serum to be 400. The rationale for this differentiation is thatcalcium, the incidence in a Japanese hospital has blood samples from MAH patients are likely to showincreased to 1:2500–5000 [9,10]. In screening for hypochloremia secondary to metabolic alkalosis,PHPT in symptomatic outpatients and those that are while blood samples from PHPT patients showhospitalized, it is difficult to differentiate PHPT from hyperchloremia due to the action of PTH. Therefore,malignancy-associated hypercalcemia (MAH) [11– blood samples in which the serum calcium, phos-13]. The differential diagnosis between these two phate, albumin and chloride concentrations had beendisorders can be achieved by detailed analysis of the measured were selected for this study. Patients withcase history and results of an assay for parathyroid blood samples that satisfied Formulae 1 and 2 werehormone (PTH). However, these procedures are considered to be positive in the screening for PHPT,time-consuming and costly, thus reducing their use- and were subject to intensive evaluation. The HISfulness for screening. In the present study, we took was programmed to detect samples satisfying For-advantage of the complete computerization of patient mula 1, and those satisfying Formulae 1 and 2.information in a new hospital information system at Patients with blood samples satisfying these formulasour hospital, and attempted to establish a highly- underwent the secondary evaluation. The screeningspecific screening method for the preliminary diag- team examined the hospital records of these patients,nosis of PHPT based on routine serum tests [14]. and if necessary, requested the physicians in charge

of outpatients and inpatients to reexamine or tomeasure the serum concentration of intact-PTH (Fig.

2. Methods and materials 1).

Blood samples that had been obtained from out-patients and inpatients at Osaka University Hospital 3. Resultsin which the serum calcium (Ca), inorganic phos-phate (IP), albumin (Alb) and chloride (Cl) con- 3.1. PHPT discovered by this screening procedurecentrations had been measured during the one-yearperiod between March 1994 and February 1995, During the study period, serum calcium, phos-were included in this study. Based on the serum phate, albumin and chloride concentrations had beencalcium and albumin concentrations, samples with a measured in 25 179 blood samples, the majoritycalcium concentration exceeding [regression (69%) of which had been obtained from inpatients.formula13S.D.] were selected as those with hy- Of the 25 179 blood samples, 824 samples obtainedpercalcemia (Formula 1 in Table 1). A blood sample from 135 patients showed hypercalcemia satisfying

Table 1aFormulae in the screening for PHPT

Formula 1 Ca 0.3 Alb 4.1Formula 2 (Cl284) (10 Alb215) (IP 3.1) 400

a Calcium (Ca), chloride (Cl); mEq/ml., phosphate (IP); mg/dl., albumin (Alb); g /dl.

S.J. Kim et al. / Clinica Chimica Acta 305 (2001) 35 –40 37

Fig. 1. Scheme of screening for PHPT. *Network of patient information in the clinical laboratory supervised system (CLASSY) [9].

Table 3Formula 1 (Table 2). Of the blood samples satisfyingCauses of the 35 false-positive cases with blood samples positiveFormula 1, 214 blood samples from 54 patients werefor Formulas 1 and 2positive in the screening for PHPT, satisfying both

aBlood purification therapy 22 casesformulas. Nineteen (35.2%) of these 54 patients hadDIC 6 casesbeen subsequently diagnosed with PHPT. Of the 19

bMalignant tumors 4 casespatients with PHPT, three had recurrent PHPT andFamilial hypocalciuric hypercalcemia 1 case

one had a malignant disease coexisting with PHPT. Administration of vitamin D 1 caseIn addition, this number includes two PHPT patients Unknown 1 case

awho were discovered by the present screening. After Includes hemodialysis and plasma exchange.bconfirming that the intact-PTH concentration of these All patients who had not received blood purification therapy

or anti-DIC therapy, and who were found to have a malignanttwo patients was high, they underwent surgicaltumor. One patient had recurrent breast cancer with PTH-relatedexcision of the parathyroid gland and had successfulprotein.outcomes.

3.2. False-positive patients received some plasma purification therapy such asacute hemodialysis, plasma exchange or plasma

Thirty-five (64.8%) of the 54 positive patients in absorption. These 22 patients showed high-gradethis study were false-positive, i.e., their blood sam- abnormality of electrolytes. As many as six casesples were positive for Formulas 1 and 2 but the were in a state of disseminated intravascular coagula-patients were subsequently found not to have PHPT. tion syndrome (DIC). Four of the 35 false-positiveThe causes of the false-positive result in the 35 cases cases had a malignancy. Therefore, this screeningare listed in Table 3. Of the 35 cases, 22 cases had revealed a 7.4% (4 of 54 patients) false-positive rate

Table 2Results of screening

aNumber of eligible blood samples for screening 25 179 samplesNumber of patients with blood samples positive for Formula 1 135 casesNumber of patients with blood samples positive for Formulas 1 and 2 54 cases

b cNumber of patients with PHPT 19 casesNumber of patients with PHPT discovered by the screening 2 cases

a 31% of the samples had been obtained from outpatients and 69% had been obtained from inpatients.b Number of patients who had been diagnosed with PHPT prior to the present study and patients diagnosed with PHPT in the present

screening.c Includes three recurrent PHPT cases and one PHPT case coexisting with malignancy.

38 S.J. Kim et al. / Clinica Chimica Acta 305 (2001) 35 –40

Table 4 diagnosed with MAH, and 27 patients (33%) hadCauses of hypercalcemia in the 81 hypercalcemic patients who been diagnosed with a severe congenital disease orwere negative for Formula 2

pediatric disease. Of the 33 patients with MAH inMalignant tumors (MAH) 29 cases (36%) the present study, the percentage of patients withPediatric diseases 27 cases (33%) each primary cancer is shown in Table 5.aBlood purification therapy 13 cases (16%)Orthopedic diseases 5 cases (6%)Others 7 cases (9%) 3.4. False-negative patient

a Includes hemodialysis and plasma exchange.

To assess false negatives, the HIS searched forpatients with PHPT. Twenty patients were detected

caused by MAH. Of these 4 cases, one had recurrentin this manner and were eligible for this study.

breast cancer that was producing a PTH-relatedAlthough 19 of these 20 patients were found to have

protein (PTHrP). The other three cases had advancedPHPT by this screening, only one patient had not

or terminal cancer of the lung, esophagus and uterinebeen discovered because the blood samples from that

cervix, respectively. The PTHrP concentration waspatient had satisfied neither Formula 1 nor 2 during

not measured in these three, who histopathologicallythe study period. The accuracy diagnosis of the

had squamous cell cancers. As to the other false-false-negative patient was a normocalcemic PHPT.

positive cases, one had familial hypocalciuric hy-percalcemia, and one was taking oral vitamin D. Ofthe 35 false-positive cases, nine cases were beingtreated in the Department of Pediatrics or Pediatric 4. DiscussionSurgery, eight in the ICU, seven in the Departmentof Surgery, five in the Department of Urology, and With the advent of the automated chemistrysix in other departments. Approximately one-half of analyzers, measuring the serum calcium concentra-the false-positive patients were being treated in either tion has become easy, and consequently, the prob-the Department of Pediatrics or ICU. ability of detecting PHPT has considerably increased

[1–13]. Heath et al. [3] reported that the incidence of3.3. Cause of hypercalcemia PHPT in the general population was

7.861.2:100 000 prior to the introduction of routineThe blood samples of 81 patients with hypercal- measurement of serum calcium concentration, while

cemia satisfied Formula 1 but not Formula 2 (Table it was 51.169.6:100 000 after its introduction. In4). Of these 81 patients, 29 patients (36%) had been Japan, the incidence of PHPT was reported to be

Table 5aPrimary diseases of malignancy-associated hypercalcemia

Present study Previous studies(n533)% of cases Lind and Ljunghall Strewler et al.

(n553) (n5444)% of cases % of cases

Lung cancer 16 29 25.0Renal cell cancer 9 20 7.9Breast cancer 13 12 19.6Hematologic cancer 9 14 12.9Prostatic cancer 3 6 –Esophageal cancer 13 – 5.6Others 37 19 29.0

a Percentage of patients with MAH who had the indicated primary cancer, in the present study and in the studies of Lind and Ljunghall[13] and Strewler et al. [18].

S.J. Kim et al. / Clinica Chimica Acta 305 (2001) 35 –40 39

1:2500–5000 after the introduction of screening by PTHrP. The present method is excellent for themeasurement of serum calcium [10]. purpose of differentiating between PHPT and MAH.

However, although the serum calcium concen- However, it is based on the serum concentrations oftration has high sensitivity for PHPT screening, it chloride and phosphate. This method is likely to givealso has low specificity. False-positivity is always a a false-positive result in patients who have received aproblem, particularly for symptomatic outpatients blood purification therapy such as hemodialysis orand hospitalized patients, as it cannot accurately plasma exchange. Care must be taken in selectingdifferentiate between PHPT and MAH. Lafferty et al. subjects suitable for the screening.[12] reported that 54–65% of the patients with MAH is present in 5–10% of all cancers [17–19].hypercalcemia in a hospital had MAH, while only The complication rate of MAH among terminal21–27% had PHPT. Similarly in the study by Walls cancers is much higher than that among non-terminalet al. [15], hypercalcemia was attributed to hyper- cancers. Adult T-cell leukemia (ATL) has the highestparathyroidism (HPT) in 56%, MAH in 29%, and complication rate of MAH among all cancers ofHPT coexisting with malignancy in 11% of cases in 60–80%. ATL is rare in the US and is endemic ina district general hospital, while in a regional oncolo- Japan [18,19]. Therefore, a large percentage of lunggy center, hypercalcemia was attributed to malig- cancer, breast cancer, and multiple myeloma casesnancy in 96% and HPT coexisting with malignancy have MAH, and these cases comprise more than 50%in 4%. Therefore, MAH predominated among the of all cases of MAH [13,17–20]. MAH is assumed todiseases of patients with hypercalcemia treated in a be a rather common finding in routine blood exami-hospital, especially in those that specialize in malig- nation, but physicians sometimes have not been ablenant disorders [16]. to detect MAH because these patients have various

PHPT can easily be differentiated from MAH by nonspecific symptoms. MAH is commonly found indetailed examination of the patient’s clinical records cancers of the lung, esophagus, kidney, urinaryand by measuring the intact-PTH concentration [15]. bladder and mammary gland, and in hemopoieticHowever, these procedures are time-consuming and malignancies [13,18]. This finding was also con-it is costly to carefully analyze the clinical records of firmed in the present study (Table 5).many patients. Lind and Ljunghall [13] noticed that Most of MAH are caused by PTHrP, and they arePTH promotes hyperchloremic metabolic acidosis by so-called humoral hypercalcemia of malignancy

1 1inhibiting Na –H exchange and increasing bicar- (HHM) [19,21]. Today, it is well known that PTHbonate excretion in the tubule, while MAH promotes and PTHrP react with the same receptor and PTHrPhypochloremia secondary to metabolic alkalosis. mimics the actions of PTH in vivo and vitro [22–They used the serum chloride concentration alone to 24]. However, prolonged action of PTHrP is reporteddifferentiate between PHPT and MAH. Setting the to restrict bicarbonate excretion compared to that ofborderline value at Cl5104 mEq/ml, the serum PTH. This difference in reactivity to bicarbonatechloride concentration can correctly diagnose 92% of excretion might produce a hyperchloremic acidosisMAH patients and 86% of PHPT patients. They in PHPT and a metabolic alkalosis in MAH [25]. Inadded the serum albumin and phosphate concen- addition, patients with MAH often have dehydrationtrations to produce Formula 2 used in the present and malnutrition, which leads to metabolic alkalosis.study. We partially modified their method and per- The purpose of the present study was to examineformed the present screening. Of the patients with whether it is possible to differentiate PHPT andhypercalcemia in the present study, 33 patients had MAH based on routine blood tests. Therefore, bloodMAH. The blood samples of 29 MAH patients samples in which serum calcium, phosphate, albuminsatisfied only Formula 1, and the blood samples of 4 and chloride concentrations had been measured wereMAH patients satisfied Formulas 1 and 2. Of these selectively included in this study. The subjects werefour cases, one case actually showed a high PTHrP patients seen at a university hospital, the majority ofconcentration, and the tumors of the other three whom were inpatients. The sampling appears to bepatients which were histologically proven to be biased. To further establish the usefulness of thesquamous cell cancers, might also have secreted present screening method, it is necessary to study as

40 S.J. Kim et al. / Clinica Chimica Acta 305 (2001) 35 –40

[10] Nakazawa H, Fujimoto T, Akiyama H, Shishiba Y,many blood samples as possible, to limit the ageNakayama T. Primary hyperparathyroidism – incidence andrange of the subjects, and to exclude blood samplesdiagnosis (in Japanese). Endocrine Surgery 1984;1:149–55.

that had been obtained from pediatric patients and [11] Harrop JS, Bailey JE, Woodhead JS. Incidence of hypercal-those in the ICU. caemia and primary hyperparathyroidism in relation to the

biochemical profile. J Clin Pathol 1982;35:395–400.[12] Lafferty FW. Differential diagnosis of hypercalcemia. J Bone

Miner Res 1991;6(suppl 2):51–9.Acknowledgements[13] Lind L, Ljunghall S. Serum chloride in the differential

diagnosis of hypercalcemia. Exp Clin EndocrinolThis study was supported in part by a Grant-in- 1991;98:179–84.

Aid for Scientific Research (11557085) from the [14] Maeda I, Hayashi S, Takeoka K et al. Development of aclinical laboratory supervised system (CLASSY). Med In-Ministry of Education, Science, and Culture ofform 1995;20:35–43.Japan.

[15] Walls J, Ratcliffe WA, Howell A, Bundred NJ. Parathyroidhormone and parathyroid hormone-related protein in theinvestigation of hypercalcaemia in two hospital populations.

References Clin Endocrinol 1994;41:407–13.[16] Fisken RA, Heath DA, Bold AM. Hypercalcaemia – A

hospital survey. Q J Med 1980;49:405–15.[1] Boonstra CE, Jackson CE. Hyperparathyroidism detected by[17] Strewler GJ, Nissenson RA. Hypercalcemia in malignancy.routine serum calcium analysis. Ann Intern Med

West J Med 1990;153:635–40.1965;63:468–74.[18] Strewler GJ, Nissenson RA. Nonparathyroid hypercalcemia.[2] Christensson T, Hellstrom K, Wengle B, Alveryd A, Wikland

Adv Intern Med 1987;32:235–58.B. Prevalence of hypercalcaemia in a health screening in[19] Strewler GJ. Hypercalcemia of malignancy and parathyroidStockholm. Acta Med Scand 1976;200:131–7.

hormone-related protein. In: Clark OH, Duh Q-Y, editors,[3] Heath H, Hodgson SF, Kennedy MA. Primary hyperpara-Textbook of endocrine surgery, Philadelphia: WB Saunders,thyroidism: incidence, morbidity, and potential economic1997, pp. 426–31.impact in a community. N Engl J Med 1980;302:189–93.

[20] Wysolmerski JJ, Broadus AE. Hypercalcemia of malignancy:[4] Mundy GR, Cove DH, Fisken R. Primary hyperparathyro-the central role of parathyroid hormone-related protein. Annuidism: changes in the pattern of clinical presentation. LancetRev Med 1994;45:189–200.1980;1:1317–20.

[21] Broadus AE, Mangin M, Ikeda K et al. Humoral hypercal-[5] Consensus development conference statement. J Bone Minercemia of cancer. Identification of a novel parathyroid hor-Res 1991;6(suppl 2):9–13.mone-like peptide. N Engl J Med 1988;319:556–63.[6] Palmer M, Jakobsson S, Akerstrom G, Ljunghall S. Preval-

[22] Kemp BE, Moseley JM, Rodda CP et al. Parathyroidence of hypercalcaemia in a health survey: a 14-year follow-hormone-related protein of malignancy: active syntheticup study of serum calcium values. Eur J Clin Invetfragments. Science 1987;238:1568–70.1988;18:39–46.

[23] Horiuchi N, Caulfield MP, Fisher JE et al. Similarity of[7] Lindstedt G, Nystrom E, Lundberg P, Johansson E, Egger-synthetic peptide from human tumor to parathyroid hormonetsen R. Screening of an elderly population in primary carein vivo and in vitro. Science 1987;238:1566–8.for primary hyperparathyroidism. Scand J Prim Health Care

[24] Rodan SB, Noda M, Wesolowski G, Rosenblatt M, Rodan1992;10:192–7.GA. Comparison of postreceptor effects of 1-34 human[8] Lundgren E, Rastad J, Thurfjell E, Akerstrom G, Ljunghallhypercalcemic factor and 1-34 human parathyroid hormoneS. Population-based screening for primary hyperparathyro-in rat osteosarcoma cells. J Clin Invest 1988;81:924–7.idism with serum calcium and parathyroid hormone values in

[25] Ellis AG, Adam WR, Martin TJ. Comparison of the effectsmenopausal women. Surgery 1997;121:287–94.of parathyroid hormone (PTH) and recombinant PTH-related[9] Kobayashi T, Sugimoto T, Chihara K. Clinical and bio-protein on bicarbonate excretion by the isolated perfused ratchemical presentation of primary hyperparathyroidism inkidney. J Endocrinol 1990;126:403–8.Kansai District of Japan. Endocrine J 1997;44:595–601.