Postoperative hungry bone syndrome in primary hyperparathyroidism: risk factors and outcomes at a South African tertiary centre

Abstract

Background

Hungry bone syndrome (HBS) is a complication after parathyroidectomy (PTX) for primary hyperparathyroidism (PHPT), with variable prevalence worldwide. Data on predictors and outcomes are limited. This study assessed the prevalence, risk factors, and outcomes of HBS after PTX for PHPT.

Methods

A retrospective review of patients undergoing PTX for PHPT (2003–2023) at Inkosi Albert Luthuli Central Hospital, KwaZulu-Natal, South Africa. HBS was defined as hypocalcaemia (corrected serum calcium < 2.1 mmol/l) lasting > 4 days post-PTX with normal or elevated intact parathyroid hormone (iPTH). Clinical, biochemical, radiological data and hospital stay were recorded.

Results

Of 117 patients (86.4% female; 71,8% symptomatic) included in this analysis, HBS developed in 20.4% (n:24). Compared with the no-HBS group (n = 93), HBS patients were younger, leaner, and more likely to have osteoporosis or osteitis fibrosa cystica (OFC). They had higher preoperative iPTH (162.8 [104–208.7] vs. 23.1 [15.6–37.5] pmol/L, P < 0.001) and ALP (696 [308.5–1323.5] vs. 111.5 [88.0–134.2] U/L, P < 0.0001), with lower 25-OH vitamin D. Hospital stay was longer (19.5 [11.0–31.2] vs. 4.0 [3.0–4.0] days, P < 0.0001). ROC analysis identified iPTH > 121 pmol/L (Sn 73.9%, Sp 100%) and ALP > 302 IU/L (Sn 78.3%, Sp 98.9%) as cut-points for HBS. In multivariate analysis, independent predictors were OFC (OR 7.7, 95%CI 1.5–40.6; P = 0.016) and iPTH > 121 pmol/L (OR 100.5, 95%CI 4.6–2225.8; P = 0.004).

Conclusion

HBS remains a relevant postoperative complication in PHPT, particularly in patients with advanced skeletal disease. Elevated preoperative iPTH levels and OFC may serve as useful markers for risk stratification. These indicators can assist in guiding perioperative management and optimising patient care. HBS prolonged the hospital stay, highlighting its clinical significance and implications for healthcare resource utilisation. Early identification of high-risk patients is therefore crucial to inform perioperative strategies and improve surgical outcomes.

Clinical trial number

Not applicable.

Introduction

Primary hyperparathyroidism (PHPT) is a common endocrine disorder with a presentation that has evolved over time from being severe and symptomatic in the past, to asymptomatic and/or incidentally discovered more recently [1]. The majority of western countries report a predominantly asymptomatic spectrum of disease [1]. By contrast, many developing nations continue to encounter a more extreme form of PHPT [2–4]. Notwithstanding these differences, parathyroidectomy (PTX) has universally been shown to have high success rates for cure and remains the definitive therapy for PHPT [1]. Surgical complication rates from PTX such as bleeding, infection and recurrent laryngeal nerve injury are lower in high-volume centres [5]. However, some complications such as Hungry bone syndrome (HBS) may be unavoidable despite meticulous surgical technique and successful surgery [6].

HBS typically develops post-PTX and is defined as hypocalcaemia that is profound and prolonged for more than 4 days, often in conjunction with hypophosphataemia or hypomagnesaemia [6, 7]. HBS can present within hours after parathyroid surgery for PHPT with severe hypocalcaemia. The pathogenesis is thought to be related to the sudden shift in bone remodelling from a predominantly osteoclastic to an osteoblastic state, due to the sudden reduction in parathyroid (PTH) levels that occurs after successful surgery [6]. This rapid onset of osteoblastic activity leads to movement of serum calcium into the bone resulting in hypocalcaemia that may persist for weeks to months.

In addition to the variability in the presentation of PHPT between high-income and low-to-middle income countries (LMIC), differences in the prevalence of HBS have also been observed (Table 1). A higher frequency of HBS is often reported from low and middle income countries such as India, China, Egypt and Thailand compared to the rare occurrence in high-income countries like Austria, United States and Singapore [3, 4, 8–19]. The exact aetiology for this dissimilarity remains unclear but is most likely related to the frequency of severe disease, particularly skeletal manifestations [6]. Preoperative predictors of HBS reported in the literature include older age, higher iPTH, ALP and calcium as well as radiological evidence of bone involvement [20].

Table 1.

Reported prevalence of hungry bone syndrome by world bank country income classification

High Income Countries				Middle or low income countries
Country (Author)	Year	Ref.	%	Country (Author)	Year	Ref.	%
Singapore (Chandran)	2022	[17]	2.4	Turkey (Kaya)	2016	[15]	13.4
Saudi Arabia (Malabu)	2007	[18]	4.0	India (Mayilvaganan)	2017	[16]	15.8
USA (Tang)	2022	[19]	7.0	South Africa (Govind)	2024	[21]	16.7
Austria (Kaderli)	2018	[8]	8.6	Iran (Hadi)	2025	[12]	17.3
Portugal (Ferreira)	2021	[9]	9.8	Egypt (Salman)	2021	[13]	17.6
Spain (Martinez)	2020	[10]	12.2	Algeria (Zitouni)	2021	[4]	19.3
Lithuania (Jakubauskas)	2018	[14]	14.8	Thailand (Prasarttong-osoth)	2012	[3]	22.0
				Phillipines(Padilla-Barouidan)	2018	[11]	35.0

Ref: Reference

There are no established guidelines for the management of HBS, which require intensive in-hospital management. Treatment entails restoring eucalcaemia, initially using intravenous calcium, followed by high doses of oral supplementation and the addition of oral vitamin D analogues [6]. Although a recent study from South Africa reported an increase in asymptomatic PHPT, a notable proportion of patients (16.7%) developed HBS post-PTX [21]. It would be ideal if a scoring system could be developed to predict HBS. This may facilitate the institution of a preoperative optimization protocol to mitigate or even prevent HBS. Retrospective data from small case series and case reports of PHPT suggest some benefit in the use of preoperative bisphosphonate to mitigate HBS [20]. The aim of this study was to determine the burden of HBS over the past 20 years at Inkosi Albert Luthuli Central hospital (IALCH), a tertiary referral hospital in South Africa, and to determine the risk factors for the development of HBS in patients undergoing surgery for PHPT.

Research design and methods

Study design and study population

This was a retrospective analysis of all consecutive patients with PHPT who underwent PTX between 2003 and 2023 at IALCH in KwaZulu-Natal, South Africa. No patients were excluded for missing clinical or laboratory data. HBS was defined as hypocalcaemia that was profound (corrected serum calcium < 2.1 mmol/l) and prolonged for more than 4 days following PTX, with normal or raised serum intact parathyroid hormone (iPTH). Patients with familial hypocalciuric hypercalcaemia, tertiary hyperparathyroidism, under 12 years of age, lithium and diuretic-associated hypercalcaemia were excluded.

Data collection

The files of all patients with PHPT were accessed using the electronic health record (EHR) system at IALCH for clinical, radiological and laboratory information. Data recorded for each patient included demographic details, symptomatology, past medical history and clinical characteristics. Pre-operative laboratory results recorded included serum iPTH, corrected calcium, phosphate, magnesium, urea, creatinine, total alkaline phosphatase (ALP, measured using the hospital laboratory’s standard assay), 25-hydroxyvitamin D, lipid profile and HbA1c. Estimated glomerular filtration rate (eGFR) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. Twenty-four hour urine calcium excretion and spot urine calcium to creatinine excretion ratio were also recorded. PTH was measured by chemiluminescence (Centaur XP pre-2019, Atellica IM 2019–present; reference range 2.0–8.5 pmol/L), and 25‑OH‑Vitamin D by HPLC (pre-2016) then LC–MS, with deficiency defined as < 30 nmol/L.

Osteoporosis was defined as a bone mineral density (BMD) T-score ≤ -2.5 below the young adult mean or Z-score ≤ -2.5 below the mean for males < 50 years and premenopausal females, at any site. Pre-operative radiological data recorded included skeletal survey, bone mineral density (BMD) scan using dual-energy X-ray absorptiometry (DEXA), renal ultrasound and technetium (Tc-99 m) sestamibi scan. Osteitis fibrosa cystica (OFC) was defined as the presence of one or more of the following radiologic features: subperiosteal bone resorption, distal clavicle tapering, skull “salt-and-pepper” appearance, or lytic bone lesions consistent with brown tumours. Histology reports were reviewed for a final diagnosis (parathyroid adenoma, hyperplasia or carcinoma).

Preoperative flexible laryngoscopy was routinely performed to document baseline vocal cord function. Intraoperative nerve monitoring was not routinely used, and postoperative laryngeal function was assessed clinically for voice change; postoperative laryngoscopy was performed only when clinically indicated.

IALCH post-operative management protocol for PHPT: Serum calcium, phosphate, magnesium, and albumin were measured 6-hourly for the first 48 h and then less frequently until discharge. Serum iPTH and ALP were measured in the first 24 h. Patients were monitored for symptoms of hypocalcaemia. Asymptomatic or mild hypocalcaemia (adjusted calcium ≥ 1.9 mmol/L) was treated with oral calcium carbonate (500 mg twice daily, titrated up to 1000 mg every 8 h) to maintain calcium in the low-normal range; a vitamin D analogue (1-alpha-calcidol 0.5–1.0 µg daily) was added if hypocalcaemia persisted. Symptomatic hypocalcaemia, adjusted calcium < 1.9 mmol/L, or ECG changes prompted intravenous calcium gluconate, administered as an initial bolus (10 ml of 10% calcium gluconate over 15 min), followed by a continuous infusion (0.5–1.5 mg elemental calcium/kg/hr; prepared by adding 100 ml of 10% calcium gluconate to 1 L of 0.9% saline or 5% dextrose, initiated at 40 ml/hour and titrated to maintain serum calcium between 2.0 and 2.25 mmol/L). Oral calcium and 1-alpha-calcidol were started as soon as oral intake was tolerated. Magnesium deficiency was corrected as needed, and patients were monitored for vocal cord paresis and haematoma. Follow-up was scheduled 2–4 weeks after discharge.

Statistical analysis

Statistical analysis was performed using SPSS (Statistics Package for the Social Sciences) version 28 and Stata 17 (StataCorp LLC, USA). Continuous variables are expressed as mean (± SD) or median (IQR) and categorical variables as percentages. Univariate analysis was performed to compare differences between those with and without HBS. Spearman’s correlation analysis was used to assess linear relationships between continuous variables. Optimal cut points for PTH and ALP were investigated using receiver operating curve characteristic (ROC) analysis and the highest Youden’s index was used to determine the optimal preoperative cut points for predicting HBS. Risk factors for HBS were determined with multivariate analysis using Firth’s logistic regression analysis with 6 independent variables preselected using clinical criteria and included: pre-operative ALP, OFC, pre-operative PTH, corrected calcium > 3 mmol/L, osteoporosis, and 25-hydroxyvitamin D. For categorical variables, missing values were retained by creating a separate “missing” category to preserve sample size in multivariate models. No data imputation was performed. A manual backward elimination approach was applied, and model comparison was conducted using likelihood ratio tests to derive the final model. Only variables that remained statistically significant in the final model were retained. Adjusted odds ratios and 95% confidence intervals were reported. A P value of < 0.05 was considered statistically significant.

Results

Of 159 patients who were diagnosed with PHPT, 117 (73.6%) had surgery and 42 (26.4%) were managed medically. For the purposes of this analysis, the results presented are those for patients who had parathyroidectomy, defined as the total study group (n: 117) and when divided by HBS status, as HBS (n: 24) and no HBS (n: 93) (Tables 2, 3 and 4).

Table 2.

Preoperative clinical and biochemical characteristics in the total study group (n:117) and by hungry bone syndrome (HBS) status

	Total Group (n:117)	HBS (N:24)	No HBS (N:93)	P
Sex				< 0.01
Female	99 (84.6)	15 (62.5)	84 (90.3)
Male	18 (15.4)	9 (37.5)	9 (9.7)
Ethnicity*
African	50 (42.7)	17 (70.8)	33 (35.5)	-
Indian	49 (41.9)	5 (20.8)	44 (47.3)	< 0.01
White	16 (13.7)	1 (4.2)	15 (16.1)	0.062
Mixed Race	2 (1.7)	1 (4.27)	1 (1.1)	0.586
Age (years)	56 [42;65]	37.0 [23.0;54.8]	58.0 [50.0;67.0]	< 0.001
BMI (kg/m2)	26.8 [22.7;30.9]	23.2 [17.9;27.7]	27.3 [23.7;31.2]	< 0.01
Asymptomatic	33 (28.2)	4 (16.7)	29 (31.2)	-
Symptomatic	84 (71.8)	20 (83.3)	64 (68.8)	0.195
Constipation	35 (29.9)	8 (33.3)	27 (29.0)	0.647
Weight Loss	29 (24.8)	11 (45.8)	18 (19.4)	0.009
Headaches	21 (17.9)	2 (8.3)	19 (20.4	0.230
Nephrolithiasis	21 (17.9)	3 (12.5)	18 (19.4)	0.518
Abdominal Pain	27 (23.1)	7 (29.2)	20 (21.5)	0.393
Bone Pain	38 (32.5)	13 (54.2	25 (26.9)	0.013
Dyspepsia	22 (18.8)	4 (16.7)	18 (19.4)	0.850
Nausea/Vomiting	11 (9.4)	4 (16.7)	7 (7.5)	0.150
Memory affected	13 (11.1)	0 (0)	13 (14.0)	0.149
Depression	10 (8.6)	0 (0)	10 (10.8)	0.215
Haematuria	2 (1.7)	1 (4.2)	1 (1.1)	0.245
Psychosis	1 (0.9)	0 (0)	1 (1.1)	0.889
Biochemistry [Median (IQR)]
Calcium (mmol/l)	2.9 [2.7;3.1]	3.1 [2.9;3.3]	2.9 [2.7;3.1]	0.098
Magnesium (mmol/l)	0.8 [0.7;0.9]	0.8 [0.7;0.8]	0.8 [0.8;0.9]	0.285
Phosphate (mmol/l)	0.9 [0.7;1.0]	0.8 [0.6;0.9]	0.9 [0.7;1.0]	0.032
Creatinine (µmol/l)	70.5 [57.2;94.8]	61.0 [37;73.5]	72.0 [58.5;96.0]	0.157
eGFR (ml/min/1.73m2)‡	87 [59.5;110]	107.0 [81;147]	82.0 [56.2;102.2]	< 0.01
ALP (U/L)	118 [91;175]	696[308;1323]	111.5 [88;134.2]	< 0.001
iPTH (pmol/l)	26 [16.6;58.7]	162.8 [104;208.7]	23.1 [15.6;37.5]	< 0.001
25-OH Vit D (nmol/l)	39.7 [31;56]	35.6 [24.1;39.9]	42.5 [32.2;60.3]	< 0.05
Urine calcium (mmol/d)	3.5 [1.9;6.7]	6.1 [2.2;8.4]	3.4 [1.8;5.7]	< 0.01

Data are presented as N (%) or Median [IQR]; BMI: body mass index; ^‡Higher eGFR in the HBS group despite lower creatinine is likely due to younger age and lower BMI, suggesting lower muscle mass; P value: HBS vs. no HBS; *P-values for ethnicity represent pairwise comparisons of African patients versus each of the other ethnic groups. Reference ranges: Corrected calcium: 2.15–2.50 mmol/L; Alkaline phosphatase (ALP): 42–98 U/L; Intact parathyroid hormone (iPTH): 2.0–8.5 pmol/L; 25-hydroxyvitamin D (25-OH Vit D): deficiency < 30 nmol/L

Table 3.

Preoperative radiologic features and bone mineral density (BMD) findings in the total study group and in patients with and without hungry bone syndrome (HBS) (n:117)

	Total Group (N:117)	HBS (N:24)	No HBS (N:93)	P
Radiological Features*
Fractures	14 (12.0)	12 (52.2)	2 (2.4)	< 0.001
Subperiosteal resorption	25 (21.4)	15 (65.2)	8 (9.5)	< 0.001
Brown tumours	12 (10.3)	11 (47.8)	1 (1.2)	< 0.001
Bone cysts	13 (11.1)	11 (47.8)	2 (2.4)	< 0.001
Osteitis fibrosa cystica	28 (23.9)	20 (87)	8 (9.5)	< 0.001
Bone Mineral Density
Hip aBMD (g/cm2)	0.8 [0.6;0.9]	0.5 [0.3;0.7]	0.8 [0.7;0.9]	< 0.001
Spine aBMD (g/cm2)	0.8 [0.7;0.9]	0.6 [0.5;0.7]	0.8 [0.7;0.9]	< 0.001
Radius aBMD (g/cm2)	0.5 [0.4;0.6]	0.4 [0.3;0.5]	0.5 [0.5;0.6]	< 0.001
Hip T score	-1.4 [-2.6;-0.7]	-3.4 [-5.3;-2.7]	-1.2 [-2.2;-0.6]	< 0.001
Spine T score	-2.6 [-3.7;-1.5]	-4.9 [-5.4;-3.4]	-2.4 [-3.2;-1.5]	< 0.001
Radius T score	-2.7 [-4.0;-1.7]	-5.8 [-7.5;-5.1]	-2.4 [-3.3;-1.6]	< 0.001
Osteoporosis N (%)‡	68(64.7)	21(91.3)	47(57.3)	< 0.01
IV Bisphosphonate use	46 (100)	16 (66.7)	30 (32.3)	< 0.01

Data are presented as N (%) or Median [IQR]; aBMD: areal bone mineral density; IV Bisphosphonate use: preoperative intravenous bisphosphonate use; P value: HBS vs. no HBS

*Radiologic data missing for 1 patient in HBS group (N:23) and for 9 patients in No HBS group (N:84); ^‡Osteoporosis data missing for 1 patient in HBS group (N:23) and 11 patients in No HBS group (N:82)

Table 4.

Postoperative biochemistry, frequency of post-operative complications and treatment in patients with and without hungry bone syndrome (N:117)

	HBS (N:24)	No HBS (93)	P
Biochemistry
Calcium (mmol/l)	2.1 [2.0;2.2]	2.3 [2.2;2.4]	< 0.001
ALP (U/l)	805.5 [391.2;1199.8]	96.0 [79.0;129.5]	< 0.01
PTH (pmol/l)	3.5 [1.4;17.0]	4.4 [2.3;7.4]	0.267
Creatinine (µmol/l)	59.0 [45.0;82.0]	74.0 [61.5;89.0]	0.864
25-OH Vit D (nmol/l)	58.6 [41.7;80.1]	57.9 [48.6;74.1]	0.758
Pathology
Adenoma weight (grams)	2.1 [2.0;2.2]	2.3 [2.2;2.4]	0.391
Complications
Laryngeal nerve injury	0	0
Bleeding	0	0
Infection	0	0
Treatment
IV calcium [n (%)]	22 (91.67%)	3 (3.23%)	< 0.0001
Duration IV calcium (days)	8.8 (± 8.4)	0.2 (± 1.9)	< 0.001
Oral Calcium [n(%)]	23 (95.83%)‡	38 (40.86%)	< 0.001
Alphacalcidol [n (%)]	21 (87.50%)*	4 (4.30%)	< 0.0001
Length hospital stay (days)	19.5 [11.0;31.2]	4.0 [3.0;4.0]	< 0.0001

Data are presented as N (%), Median [IQR] or mean (± SD); IV calcium: intravenous calcium; HBS: hungry Bone Syndrome; P value: HBS vs. no HBS; *Alphacalcidiol dose range in-hospital: 0.5–8.0 ug daily and median dose at discharge: 2.0 ug [1.0–4.0] daily; ^‡Oral calcium median dose at discharge: 3.0 g [2.0–3.0] daily

Preoperative/Baseline

Demography and clinical presentation

In the total study group, the median age was 56 (42; 65) years; 84.6% were female and the majority were of African or Indian ethnicity (Table 2). The median BMI was 26.8 (22.7; 30.9) kg/m². At presentation, 28.2% (n:33) were asymptomatic while 71.8% (n:84) were symptomatic; the most common symptoms were constipation (29.9%) and weight loss (24.8%) (Table 2). When compared to the no HBS group, patients with HBS were significantly younger (37 vs. 58 years) (P < 0.001) and leaner (median BMI 23.2 vs. 27.3 kg/m²) (P = 0.006) and had a higher prevalence of bone pain (54.2% vs. 26.9%) (Table 2).

Biochemistry

Table 2 shows the baseline biochemistry results. In the total group, median serum calcium, ALP and PTH were high, with normal phosphate. Median 24-hour urine calcium excretion was also elevated and 25-hydroxyvitamin D was low. Compared to the no HBS group, the group with HBS had significantly higher baseline median PTH (162.8 vs. 23.1 pmol/l)(P < 0.001), ALP (696 vs. 111.5 U/L)(p < 0.001), 24-hour urine calcium (6.1 vs. 3.4 mmol/d) (P = 0.009) and eGFR (107 vs. 82 mls/min/1.73m²)) (P = 0.001) accompanied by lower creatinine, with significantly lower serum phosphate and 25-hydroxyvitamin D (35.6 vs. 42.5 nmol/l) (P = 0.011) (Table 2).

Radiology

Table 3 shows the baseline radiological features. The most frequently reported abnormality on x-ray in both the total study group and among those with HBS, was osteitis fibrosa cystica. Osteoporosis was found in 64.8% (68/105), with median T-score − 2.7 (-4.0; -1.7) at the distal one-third of radius. The magnitude of BMD reduction was greatest at the distal radius, both in the overall study group and among patients with HBS. Compared to the no HBS group, patients with HBS had lower median bone mass at all sites as well as significantly higher rates of osteoporosis (91.3% vs. 57.3%) (P = 0.008) and use of preoperative intravenous bisphosphonate therapy (66.7% vs. 32.3%) (P = 0.003) (Table 3).

Post-operative

Post-operatively (Table 4), when compared to those without HBS, patients with HBS had lower median corrected serum calcium at discharge (2.1 vs. 2.3 mmol/l) (P = 0.0003) and significantly higher median ALP levels (805.5 vs. 96.0 U/L) (P = 0.0012). Patients with HBS had a significantly higher use (91.67% vs. 3.23%) (P < 0.0001) and median duration of intravenous calcium replacement (8.8 vs. 0.2 days) (P = 0.0001). They also required a significantly greater median length of hospital stay (19.5 vs. 4.0 days) (P < 0.0001) and higher use of 1-alphacalcidol as inpatients (87.5% vs. 4.30%) (P = 0.0001).

There was significant positive correlation between duration of intravenous (IV) calcium use and post-operative length of stay (r = 0.653; P < 0.001), and pre-operative ALP (r = 0.468; P < 0.001) and PTH levels (r = 0.540; P < 0.001). There was also significant inverse correlation between duration of IV calcium use and lower bone mass at the hip (r = -0.418; P < 0.001), and spine (r = -0.367; P < 0.001) (Table 5).

Table 5.

Correlation between duration of intravenous calcium therapy and clinical or biochemical parameters

Variable	N	Spearman’s rho	P
Post-operative length of hospital stay	114	0.653	< 0.001
Pre-operative ALP	111	0.468	< 0.001
Pre-operative PTH	114	0.540	< 0.001
Pre-operative Calcium	115	0.130	0.165
1-Alphacalcidol daily dose at D/C	24	0.394	0.057
Pre-operative Hip aBMD	98	-0.418	< 0.001
Pre-operative Spine aBMD	98	-0.367	< 0.001

ALP: alkaline phosphate; iPTH: intact parathyroid hormone; D/C: discharge; aBMD: areal BMD

Using ROC curve analysis, for the development of HBS, the optimal cut-point of PTH was 121.05 pmol/l (Sn 73.9%; Sp 100%) and of ALP was 302 U/L (Sn 78.3%; Sp 98.9%) (Fig. 1).

Fig. 1.

Receiver operating characteristic (ROC) curves for preoperative biochemical markers—intact parathyroid hormone (iPTH) [A] and alkaline phosphatase (ALP) [B]—and their predictive performance for the development of hungry bone syndrome (HBS) following parathyroidectomy

In multivariate analysis, the significant predictors for HBS included pre-operative PTH ≥ 121.05 pmol/l (OR 100.5; 95% CI 4.6, 2225.8; P = 0.004) and osteitis fibrosa cystica (OR 7.7; 95% CI 1.5, 40.6; P = 0.016) (Table 6).

Table 6.

Logistic regression analysis for predictors (risk factors) for hungry bone syndrome (HBS)

Variable	Adjusted OR	95% CI	P
Pre-operative iPTH ≥ 121.05 pmol/l	100.5	4.6–2225.8	0.004
Osteitis fibrosa cystica (OFC)	7.7	1.5–40.6	0.016

ALP: alkaline phosphate; iPTH: intact parathyroid hormone

Only variables retained in the final model after backward elimination are shown. Non-significant variables excluded during stepwise regression included preoperative ALP (cutpoint 302 U/L), corrected calcium > 3mmol/l, preoperative osteoporosis, and 25-hydroxyvitamin D

Discussion

This study demonstrated a high prevalence of HBS following PTX for PHPT. All patients meeting the HBS definition required treatment, indicating clinically significant hypocalcaemia. HBS was associated with markers of disease severity, notably biochemical and skeletal manifestations. Patients with HBS had higher pre-operative ALP levels, lower BMD and were more likely to receive IV bisphosphonates prior to PTX. ROC curve analysis identified PTH and ALP thresholds of 121.05 pmol/l and 302 U/L, respectively, with high specificity for predicting the development of incident HBS. Pre-operative iPTH levels and OFC were independent predictors of HBS. Furthermore, HBS was associated with prolonged hospitalisation. To our knowledge, this is the first study from Sub-Saharan Africa (SSA) to evaluate the characteristics, predictors and outcome of HBS in patients with PHPT.

The high prevalence of HBS noted in this study is consistent with reports from LMIC [11, 13, 15, 16] (Table 1). These findings are in contrast to the lower prevalence reported in most high-income countries (HIC) where the development of HBS post-PTX is relatively uncommon [8, 17–19] (Table 1). Routine calcium monitoring in most developed countries has enabled earlier detection of PHPT resulting in a predominance of asymptomatic PHPT with relatively lower rates of skeletal involvement [1]. The lower observed prevalence of HBS in HIC is therefore not unexpected when compared to countries with predominantly symptomatic presentations.

In this study the occurrence of HBS post-PTX was associated with multiple pre-operative markers of advanced hyperparathyroid bone disease, such as higher ALP, higher iPTH levels, lower BMD and the presence of OFC. These findings are consistent with several other studies from LMIC that show an association between biochemical and radiologic features of skeletal disease severity and HBS [3, 4, 11]. Moreover, in a systematic review of HBS in PHPT, Witteveen et al. reported an association of a higher prevalence of HBS with radiologic involvement [6]. In our cohort, although BMD was reduced at all sites in both the overall group and among patients who developed HBS, the distal radius (predominantly cortical bone) had the lowest median T-scores, suggesting that cortical bone underwent greater remodelling and demineralisation compared to cancellous bone (spine), consistent with the characteristic skeletal effects of PHPT. The sine qua non of HBS is an expanded bone remodeling space which is a consequence of chronic autonomous PTH secretion in PHPT. The hypothesis is that patients with advanced skeletal manifestations have increased bone turnover and greater expansion in bone remodeling space, making them more prone to HBS after the abrupt reduction in PTH levels post-successful PTX [6]. Thus, in resource-constrained environments, the recognition of OFC with plain x-rays may provide a feasible and cost-effective means of evaluating preoperative HBS risk.

In this cohort, ROC analysis identified preoperative iPTH ≥ 121.05 pmol/L and ALP ≥ 302 U/L as strong discriminators for HBS, suggesting potential value for identifying patients with more severe disease and greater skeletal involvement. The strongest independent risk factor for HBS was an iPTH level > 121 pmol/l, which, in multivariate analysis, conferred a more than 100-fold increased risk. This observation aligns with existing literature, which frequently highlights iPTH as a key risk factor for HBS, thereby underscoring the role of disease severity as a major determinant in the development of HBS [7, 8, 14, 17]. Further evidence for the value of iPTH in predicting HBS comes from studies in patients with secondary hyperparathyroidism [22]. Although PTH has both anabolic and catabolic effects on bone, osteoclastic action predominates in PHPT due to uncoupling of osteoblastic from osteoclastic action. Therefore, the greater the elevation in PTH, the more significant its contribution to the degree of bone resorption and consequently the risk of HBS. A recent study from Iran examining 25 patients with HBS, demonstrated higher pre-operative iPTH and calcium levels in affected individuals but failed to demonstrate the value of these pre-operative parameters in predicting HBS [12]. The authors emphasize the importance of post-operative biochemical monitoring; however, ALP levels, OFC and BMD were not included in the predictive model. The findings in our study argue for the potential utility of incorporating these skeletal parameters in predicting HBS, particularly in the assessment of symptomatic PHPT.

In this study ALP levels and hypercalcaemia were not found to be independent risk factors for HBS. ALP is a surrogate marker of osteoblastic activity and has been shown in other studies to be a useful predictor of HBS [6–8, 13, 17, 20]. Potential explanations for this discrepancy include confounding factors, such as concurrent vitamin D deficiency, and the comparatively stronger predictive performance of iPTH levels in this study, which may have attenuated the observed effect of ALP. The lack of an independent association between hypercalcaemia and HBS in this study is unsurprising, given that it does not consistently reflect disease severity, a finding supported by other studies.

There was significantly greater use of intravenous bisphosphonates in the HBS group compared to those who did not develop HBS. Although this study was not designed to assess bisphosphonate efficacy, it was evident that short-term IV bisphosphonates were insufficient to prevent HBS. A single infusion of zoledronate within 2 weeks of PTX, typically used as a bridge to surgery, has not been shown to prevent HBS [23]. In contrast, studies from India and Taiwan have reported a protective effect with bisphosphonate therapy [16, 24]. One potential explanation for this discrepancy is the timing of therapy relative to PTX. An extremely limited study of four patients suggests that longer-term preoperative bisphosphonate use may mitigate HBS [25]. Extending therapy beyond 4–6 weeks—allowing at least one bone remodeling cycle—may reduce osteoblastic activity more effectively than short-term use, thereby limiting the rapid skeletal calcium uptake that follows PTX.

The relationship between vitamin D deficiency and HBS remains contentious. In this study, vitamin D was associated with HBS in univariate analysis, however it was not an independent predictor. Although some authors have suggested vitamin D as a risk factor [26], this has not been borne out by the available studies [7, 8, 13]. However, interest persists in the potential of vitamin D analogues to attenuate HBS severity following PTX [27, 28]. In this study, alphacalcidol was initiated promptly at HBS diagnosis, with doses up to 8 µg daily; however, delayed response necessitated prolonged IV calcium supplementation. This lag may reflect the time required for genomic activation of intestinal calcium transport pathways, including the upregulation of calbindin and transient receptor potential vanilloid type 6 (TRPV6) [29]. Further research is required to evaluate the efficacy of pre-operative administration of vitamin D analogues at therapeutic doses in high-risk patients, with the aim of reducing IV calcium requirements and shortening hospital stay.

The higher prevalence of HBS in younger individuals in this study, while consistent with two prior reports [13, 14], contrasts with studies in which HBS was associated with older age [7, 10]. This may reflect delayed diagnosis due to non-specific symptoms such as bone pain being overlooked. Given that disease severity is a key determinant of HBS risk, younger age at presentation may indicate more advanced disease at diagnosis, supported by findings from Maldar et al. showing that symptomatic HBS is more severe in premenopausal than postmenopausal women, possibly due to increased skeletal sensitivity to elevated PTH levels [30]. Additionally, lower BMI was associated with HBS in this cohort, independent of gastrointestinal symptoms. Alternative explanations include sarcopenia related to chronic disease or disuse muscle atrophy secondary to immobilisation from fractures or chronic bone pain.

In the absence of routine calcium screening in low- and middle-income countries, delayed diagnosis of PHPT will likely continue to dominate, with severe disease remaining the prevailing clinical presentation in the developing world. HBS is associated with prolonged hospital stay, underscoring its significant clinical impact and potential to strain healthcare resources. This finding emphasizes the importance of heightened clinical vigilance to identify individuals at increased risk, enabling more effective pre- and postoperative management. In our cohort, hospital stays were longer than in many other reports, reflecting our institutional protocol of intensive postoperative calcium monitoring for at least 48 h in all patients; risk stratification could enable implementation of enhanced recovery after surgery (ERAS) protocols for low-risk patients and support early discharge, thereby improving efficiency and reducing healthcare costs.

From a clinical perspective, these results underscore the need for a cost-effective perioperative strategy in resource-limited settings, prioritising higher-risk patients—particularly those with skeletal involvement, such as low bone mass or OFC—for closer monitoring and proactive management. This approach may include preoperative optimisation of vitamin D status and prophylactic vitamin D analogues in elective surgical candidates identified as high-risk for HBS; an example of such a strategy is outlined in Table 7. Although implementation may be limited by cost in low-resource settings, focusing on clearly defined high-risk groups can enhance feasibility. To support these targeted approaches, a meta-analysis of existing data is needed to develop a universal HBS risk prediction tool. Furthermore, randomised controlled trials are required to evaluate the efficacy of preventive strategies and to determine whether ERAS protocols can similarly optimise outcomes in patients with PHPT, as demonstrated in renal hyperparathyroidism [31].

Table 7.

Hypothesis-generating framework of potential preoperative measures for the prevention of hungry bone syndrome (HBS) in high-risk surgical candidates, applicable where elective surgery can be safely delayed by six weeks

Preventive Measure	Timing	Notes / Precautions
Correct vitamin D deficiency	Commence 6 weeks preoperatively	e.g., Calciferol 50,000 IU twice weekly; Stop if calcium > 3 mmol/l
Intravenous (IV) Zoledronate	At least 6 weeks before surgery AND repeat 2 weeks prior to surgery	Single dose < 2 weeks before surgery not shown to be effective. Caution with CKD
Vitamin D analogue* (e.g., alfacalcidiol 0.5 µg daily)	5–7 days preoperatively; continue after surgery with dose titration	Alternate day calcium monitoring, discontinue if calcium > 3 mmol/L

*Initiate only if all pre-requisites are met: serum calcium < 3 mmol/L, intravenous zoledronate has been administered, and regular calcium monitoring is feasible

Note: Evidence from randomised controlled trials is lacking; Recommendations are based on small studies [16, 18, 24–26, 28]

Strengths of this study include a well-characterised cohort with comprehensive BMD and radiologic data, and one of the larger HBS cohorts analyzed in SSA and globally. Limitations of this study include its retrospective nature and the absence of bone turnover markers. Imaging was not standardized for skeletal site assessment, limiting evaluation of lesion distribution; future studies should include targeted skeletal imaging to better define anatomical patterns. The predominance of female participants may have influenced the observed OFC prevalence; future studies should explore potential hormonal influences. Postoperative BMD data was incomplete and therefore not included, limiting assessment of long-term bone recovery. Although this study highlights the spectrum of disease seen in the public sector which manages most patients in South Africa, it may not be representative of that observed in the private health-care environment where routine calcium screening is more widespread.

Conclusion

HBS remains a relevant post-operative complication, particularly in patients with advanced skeletal manifestations of PHPT. Pre-operative iPTH levels and OFC may serve as valuable markers for risk stratification. These indicators can help guide perioperative management and optimise patient care. HBS was associated with prolonged hospital stay, underscoring its significant clinical impact and potential to strain healthcare resources. These findings highlight the importance of identifying high-risk patients pre-operatively to guide peri-operative management strategies and optimise resource utilization. Awareness of the risk factors and consequences of HBS is fundamental to improving surgical outcomes. Prospective studies are needed to validate these predictors and explore the efficacy of pre-emptive interventions to prevent HBS.

Abbreviations

PHPT

Primary Hyperparathyroidism

SSA

Sub-Saharan Africa

IO-PTH

Intraoperative parathyroid hormone

IALCH

Inkosi Albert Luthuli Central Hospital

PTH

Parathyroid hormone

USA

United States of America

EHR

Electronic health record

ALP

Alkaline phosphatase

eGFR

Estimated glomerular filtration rate

CKD-EPI

Chronic Kidney Disease Epidemiology Collaboration

HBS

Hungry Bone Syndrome

BMD

Bone Mineral Density

DEXA

Dual-Energy X-ray Absorptiometry

Tc-99m

Technetium

SPSS

Statistics Package for the Social Sciences

IQR

Interquartile range

OR

Odds Ratio

95% CI

95% confidence interval

BMI

Body mass index

Author contributions

Author IP: conceived and designed the study, collected the data, performed data analysis, and drafted the initial manuscript. Author KG: assisted with data collection, data verification, assisted with drafting the manuscript. Author FP: contributed to data interpretation, literature review and editing of the draft manuscript. Author AM: contributed to data interpretation and critically revised the manuscript. All authors read and approved of the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethics approval for this retrospective study was obtained from the Biomedical Research Ethics Committee (BREC), University of KwaZulu-Natal (UKZN) (approval number: BREC/00006275/2023). The requirement for informed consent was waived by the Ethics Committee due to the retrospective nature of the study and the use of anonymized/de-identified data. All procedures involving human participants were conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.