1.1.1. Which biochemical test or combination of tests should be used for diagnosing primary hyperparathyroidism (for example, levels of parathyroid hormone, blood calcium and phosphate, alone or in combination)?

1.4.1. Included studies

This review aimed to assess the biochemical tests or combination of tests that should be used for diagnosing primary hyperparathyroidism. One study, Christensen 2008¹³ was included in the review. The study evaluated the discriminative power of calcium creatinine excretion ratio, 24-hour renal calcium/creatinine clearance ratio and 24-hour renal calcium excretion for the separation between FHH and PHPT.

No relevant diagnostic test accuracy studies of index tests: PTH, ionised calcium, phosphate (morning versus fasting versus random test), and alkaline phosphatase in people under investigation for suspected primary hyperparathyroidism were identified.

The included study is summarised in Table 2 below.

1.4.2. Excluded studies

See the excluded studies list in appendix H.

Table 2

Summary of clinical studies included in the evidence review.

Table 3

Clinical evidence summary: Receiver operating characteristic (ROC) curve analysis for discrimination between patients with familiar hypocalciuric hypercalcaemia (FHH) and patients with PHPT [modified GRADE table].

1.5.1. Included studies

No relevant health economic studies were identified.

1.5.2. Excluded studies

No health economic studies that were relevant to this question were excluded due to assessment of limited applicability or methodological limitations.

1.5.3. Unit costs

The unit costs of tests for diagnostic testing were presented to aid committee discussion.

Table 4

Cost of diagnostic testing.

1.7.1. Clinical evidence statements

One study showed that 24-hour renal calcium excretion (CE) (mmol, measured directly in the urine) had a sensitivity of 87%, and a corresponding specificity of 72.2% for discriminating between patients with familiar hypocalciuric hypercalcaemia (FHH) and patients with PHPT (n=54 FHH; n=97 PHPT; Very Low quality).

One study showed that 24-hour renal calcium/creatinine excretion ratio (CR) (mmol/mmol, cut-off point <0.52) had a sensitivity of 88.9% and a corresponding specificity of 81.4% for discriminating between patients with familiar hypocalciuric hypercalcaemia (FHH) and patients with PHPT (n=54 FHH; n=97 PHPT; Very Low quality). One study showed that calcium/creatinine clearance ratio (CCCR) (cut-off point <0.0115) had a sensitivity of 79.6% and a corresponding specificity of 87.6% for discriminating between patients with familiar hypocalciuric hypercalcaemia (FHH) and patients with PHPT (n=54 FHH; n=97 PHPT; Very Low quality).

No evidence was identified for ionised calcium, phosphate, alkaline phosphatase, vitamin D, and bone turnover markers.

1.7.2. Health economic evidence statements

No relevant economic evaluations were identified.

1.8.1. Interpreting the evidence

1.8.2. The outcomes that matter most

The committee considered the following criteria of specificity, sensitivity, positive and/or negative predictive value, or area under Receiver Operating Characteristic (ROC) curve for the index test for primary hyperparathyroidism for decision making. The committee deemed both sensitivity and specificity as equally important for decision-making.

No relevant diagnostic test accuracy studies of index tests PTH, ionised calcium, phosphate (morning versus fasting versus random test), alkaline phosphatase, vitamin D and bone turnover markers in people under investigation for suspected primary hyperparathyroidism were identified.

1.8.3.

From the AUCs it appeared that CCCR gives a marginally better discrimination between familial hypocalciuric hypercalcaemia and primary hyperparathyroidism than CR and CE. The committee acknowledged that AUCs compare test accuracy over different thresholds for positivity and provide information on overall measure of the performance of the test, but in actuality these tests are usually employed at one given threshold and therefore the sensitivity/specificity at that particular threshold are more useful markers of how good the test will be in clinical practice.

1.8.4. Cost effectiveness and resource use

No relevant economic evaluations were identified for this question. Unit costs were presented to the committee for consideration.

NHS reference costs (2015–16) lists the pathology cost for clinical biochemistry (which typically includes test for calcium) to be £1.13. A urine test (using urinalysis analyser) was estimated to be £4.08 based on prices of necessary equipment as listed in the NHS supply chain catalogue and staff time according to the Personal Social Services Research Unit (PSSRU). The cost of the test for vitamin D was estimated to be £16.50, using numbers reported in a previous study which averaged the reported price from two hospitals.

The cost of PTH was estimated using a number of values reported by committee members from their experience with laboratories. While we were unable to obtain individual prices from each lab, a range of between £6.50 and £10.20 for around 12 labs was given, with an approximate average of £8.00. The committee considered these costs and agreed they reflected those used in real practice, and that there will be some variation between practices. This is notably more expensive than that for a general clinical biochemistry test including calcium, therefore the committee considered it important that the recommendations did not lead to excessive testing for PTH.

Firstly, the committee considered it important to repeat an albumin-adjusted serum calcium test to confirm that an initial elevation in serum calcium level was repeated prior to PTH testing due to the intra-individual variability in calcium levels. As the cost of a clinical biochemistry test (including that for calcium) is relatively low, the committee considered it important that there is confirmation of hypercalcaemia with a repeat albumin-adjusted serum calcium level before a more expensive PTH test is ordered. The committee considered a repeat test for calcium could potentially be cost-saving if it lowers the number of unnecessary tests for PTH. The committee also noted that if results for repeated tests for albumin-adjusted serum calcium are consistently high, but the result for PTH test is not high, it could be an indicator for more serious conditions including cancer. Hence, the committee deemed it important to have a contemporaneous calcium test alongside the PTH test.

Secondly, the committee considered it important to have different recommendations for those with different calcium levels. The committee noted that the prevalence of primary hyperparathyroidism in those with an albumin-adjusted serum calcium level of 2.6 mmol/litre and over is high. The main differential diagnosis here is cancer and a PTH serves to make the distinction between the common cause of PTH-independent hypercalcaemia, cancer, and PHPT. The committee regarded the checking of PTH as part of current practice in all patients with sustained hypercalcaemia.

The committee discussed that the prevalence of primary hyperparathyroidism in those with an albumin-adjusted serum calcium of 2.5 mmol/litre and above is lower and therefore testing for PTH is more likely to lead to a greater proportion of unnecessary PTH testing in those who do not have primary hyperparathyroidism and hence incur a high cost. Therefore the committee considered it important that in people with albumin-adjusted serum calcium above 2.5 mmol/litre, only those with a clinical suspicion of primary hyperparathyroidism have a PTH test.

The committee also discussed that the costs incurred when a diagnosis of primary hyperparathyroidism is missed could be high if the patient experiences a clinical event, such as fragility fracture and renal stones, as a result of untreated primary hyperparathyroidism, the costs of which are much higher than those associated with diagnostic tests.

The committee discussed that current practice for diagnostic testing for primary hyperparathyroidism is widely variable. They considered that if practice for sequencing of tests for albumin-adjusted serum calcium and PTH is standardised in this way, this may present an area for cost saving.

Overall, the committee noted that excessive testing in the process of diagnosing primary hyperparathyroidism should be avoided where possible. However, this should not have precedence over the importance of achieving a timely and accurate diagnosis. Thresholds for testing should be considered alongside patient history and symptoms by the relevant healthcare provider.

1.8.5. Other factors the committee took into account

Based on their experience, the committee agreed that albumin-adjusted serum calcium level is an appropriate first-line biochemical test in those with suspected hypercalcemia. The committee recommended albumin-adjusted serum calcium measurements based on physiology. The committee was confident to recommend this test as adjusted serum calcium has physiological importance. The biological effects of calcium are mediated by free calcium that is not bound to albumin and other proteins. In clinical practice an adjustment is made for serum albumin, which is the most significant protein that calcium binds to. When calcium is bound to this protein, it does not have biological activity. It is the free and metabolically available serum calcium that has biological, physiological and clinical effects. The committee was aware that there are several equations, however each laboratory need to take into account their methods for calcium and albumin and their population mean for those values rather than adopting a ‘fixed’ equation. Laboratories should regularly review what is happening to their correction calculation.

The committee noted that albumin-adjusted serum calcium measurement could be done with or without a cuff as it would not make any difference in the values as it is albumin adjusted.

In the experience of the committee it is not necessary to measure ionised calcium. They noted that this test cannot be done in primary care and it would usually be undertaken using a blood gas analyser in hospital. The committee considered that as ionised calcium measurement is a point-of-care test it is not subject to stringent quality control like laboratory based tests. Furthermore the sample has to be handled very quickly, making it a less reliable test.

The committee discussed that there was a significant inter-individual variability in calcium levels, for example biological variation for calcium is 2.1%; and 2.2% for albumin; and the population variability for calcium is 2.5%. The committee stated that the normal reference range for serum calcium as defined by Association of Clinical Biochemistry and Laboratory Medicine is 2.2–2.6 mmol/litre. The committee discussed that repeat calcium testing was performed for the purpose of validation and also to get a contemporaneous serum calcium value along with PTH, but not for diagnostic information. The committee noted that the frequency of repeat calcium testing is context and convenience driven. How often to repeat the test depends on a number of different factors including the levels of calcium reported and symptomatology.

Based on their clinical experience, the committee recommended performing a PTH test for people with an albumin-adjusted serum calcium level repeatedly 2.6 mmol/litre or above, because they are more likely to have hypercalcaemia, which is a strong indicator of primary hyperthyroidism. The committee noted that a second measurement of albumin-adjusted serum calcium was useful, as along with intra-individual and population variability, there could be external causes such as inaccuracy and imprecision of instruments which could lead to variation in the serum calcium values.

For people with an albumin-adjusted serum calcium level repeatedly 2.5 mmol/litre or above and where clinical suspicion of hypercalcaemia is high due to symptoms the committee recommended performing a PTH test. The committee agreed that not all symptoms are specific to primary hyperparathyroidism. There is a small group of patients with primary hyperparathyroidism in whom the calcium may be within the normal range (normocalcaemia) and these patients would fall under the above category. There was recognition that normocalcaemic primary hyperparathyroidism as a diagnosis was in its relative infancy and the natural history of the disease and its optimal management is still unclear. In light of above, the committee therefore agreed that setting a threshold for PTH measurement of albumin-adjusted serum calcium level repeatedly 2.6 mmol/litre or above, or 2.5 mmol/litre or above if there is clinical suspicion of hyperparathyroidism, would identify most people with primary hyperparathyroidism. The committee however noted that the vast majority of presentations of primary hyperparathyroidism are in people with hypercalcaemia.

The committee discussed that if someone has had an incidental finding of elevated albumin-adjusted serum calcium, the albumin-adjusted serum calcium test should be repeated and if it remains elevated PTH testing should be offered. The committee recognised that repeat calcium testing will reduce the number of unnecessary PTH tests. The committee considered that repeating the calcium test is necessary due to random error or changes in the level of physiologically active calcium because of alterations in blood pH or serum albumin. The committee agreed that this test could be performed either in a primary or a secondary care setting.

The committee agreed that those patients with raised albumin-adjusted serum calcium and PTH above the mid-point of the reference range would need specialist advice regarding a likely diagnosis of primary hyperparathyroidism. The committee wanted to be permissive to allow different pathways; specialist advice in some cases will be a referral and in others it will be a telephone conversation.

The committee discussed that there is ambiguity around the PTH lower cut-off and as to when primary hyperparathyroidism is ruled out. Hence, they felt that in people with hypercalcaemia and PTH below the lower limit of the reference range, alternative causes for hypercalcaemia must be sought. The committee noted that the most common cause of hypercalcaemia with a suppressed PTH is malignancy, but other non-malignant causes such as granulomatous conditions (for example sarcoidosis), and endocrine conditions (for example thyrotoxicosis) may be involved. Thus, such patients need to be investigated and referred appropriately. The committee noted that specialist referral in this context could be an oncologist, endocrinologist, endocrine surgeon etc. The committee from their experience highlighted that there could be a small proportion of primary hyperparathyroidism patients in whom the PTH levels are in the lower part of the reference range.

The committee agreed that for people with a PTH below the midpoint of the reference range and repeated albumin-adjusted serum calcium 2.6 mmol/litre or above, specialist advice should be sought for further investigations for primary hyperparathyroidism. It was recognised that lower concentrations of PTH can be seen in rare cases of PHPT.

The committee agreed that for people with a PTH within the reference range but below the midpoint of the reference range and albumin-adjusted serum calcium less than 2.6 mmol/litre, no further investigation for PHPT is required as they are unlikely to have PHPT. The committee considered that in such cases the GP could consider further investigations for alternative diagnoses if the clinical picture suggested underlying pathology.

The committee stated that PTH testing can be done on a random sample, i.e. non-fasting and at any time of day. The committee considered that even though there is a marginal diurnal variation in PTH levels, it is not large enough to be adjusted for. It was also agreed that the PTH test does not need to be repeated prior to referral. As PTH is a relatively unstable element it is important that it needs to be taken according to the relevant laboratory collection protocols. Blood collection protocols were not prioritised as a review question. The committee are aware that there are different approaches to PTH measurement and that most laboratories do specify using an EDTA blood collection tube. However specifying the anticoagulant used is not within our scope.

The committee noted that the reference range for PTH varies from lab to lab, so numerical thresholds cannot be specified in the recommendation. The committee also noted that there was a huge inter-individual variability for this test. The committee considered that PTH testing should be done with contemporaneous albumin-adjusted serum calcium testing, as it is necessary to interpret the PTH result in the context of the albumin-adjusted serum calcium level.

The committee discussed the importance of assessment of vitamin D status in all patients with primary hyperparathyroidism. The committee discussed that vitamin D deficiency should be explored in all patients with primary hyperparathyroidism as it leads to an increase in the amount of PTH that is secreted, increase in the severity of bone disease, and could also lead to higher post-operative risk. The committee therefore agreed that measuring vitamin D and correcting any deficiency is essential in diagnosis and treatment of people with primary hyperparathyroidism, but noted that correcting the deficiency does not need to precede the diagnosis. The committee recognised the importance of correcting Vitamin D deficiency, but agreed that for some primary care providers, vitamin D testing is not available. This would slow down referrals from primary care, and hence this test should be performed in secondary care to facilitate a more timely diagnosis. The committee noted that treatment of vitamin D deficiency in patients with primary hyperparathyroidism usually has little effect on serum calcium levels but is associated with a reduction in PTH. The committee accepted that treatment for vitamin D deficiency would continue post-diagnosis of primary hyperparathyroidism to ensure vitamin D is replete in the long term.

The committee discussed that Vitamin D can affect the interpretation of the urinary calcium test, hence in people who are vitamin D deficient, the specialist should interpret the urine calcium with caution. Untreated vitamin D deficiency may cause low urine calcium excretion. Correcting any deficiency may reveal normal or even elevated urine calcium excretion. However, the likelihood of a urine calcium result being low is highly unlikely. If this unlikely result is found, it is entirely appropriate to make sure that any vitamin D deficiency has been corrected. If the vitamin D deficiency has been corrected and the urine calcium is low, the diagnosis is unlikely to be primary hyperparathyroidism. As the likelihood of urine calcium being low even in vitamin deficiency is low, the committee did not make this a major feature of the diagnostic algorithm but when urine calcium is low, rarely, there is a major focus on ensuring vitamin D repletion.

The committee discussed that the PTH and albumin-adjusted serum calcium tests are performed at first presentation; hence these tests could be performed in either primary or secondary care.

The committee discussed that in people with hypercalcaemia, when PTH is elevated or within the upper part of the reference range, primary hyperparathyroidism is the most likely diagnosis but familial hypocalciuric hypercalcaemia should be considered. In this condition, the urinary calcium/creatinine ratio is low and other members of the family may have hypercalcaemia. Though familial hypocalciuric hypercalcaemia has similar biochemical features to primary hyperparathyroidism, it generally requires no treatment and therefore it is important to exclude familial hypocalciuric hypercalcaemia prior to consideration of any treatment, particularly surgery. The committee considered that urine calcium excretion tests were important discriminatory tests and hence based on their experience and evidence from the small study agreed to recommend the following tests for discriminating primary hyperparathyroidism from familial hypocalciuric hypercalcaemia: 24-hour urinary calcium excretion, renal calcium/creatinine excretion ratio and calcium/creatinine clearance ratio. Evidence was available for these tests conducted at 24 hours. The committee acknowledged that there was no evidence for urine calcium excretion performed on a spot urine sample or at 2 hours, but they felt that these would not be different from the tests conducted at 24 hours, hence they recommended that renal calcium/creatinine excretion ratio and calcium/creatinine clearance ratio tests can be conducted on a random sample. The committee discussed the setting of the urinary calcium test. Due to the difficulties obtaining correctly timed collections and with collections being made in the incorrect container secondary care was felt to be the place for this test. Also the committee did not want to slow down referrals or discussions with secondary care. The committee discussed that the purpose of these tests is to avoid surgery in people who will not benefit from surgery. The committee agreed that the literature is inconsistent about the thresholds for these tests and hence did not recommend any specific thresholds. The committee from their experience discussed that in practice not all three tests are offered, but one of the tests (usually calcium creatinine excretion ratio) is performed.

The committee discussed that if the above calcium/creatinine ratio tests are positive then these patients would undergo further genetic testing for definitive diagnosis of familial hypocalciuric hypercalcaemia and thus spare patients with familial hypocalciuric hypercalcaemia from unnecessary surgical treatment. If these tests are false negative, then these patients would undergo unnecessary surgery and it would result in ‘failed surgery’ as these patients do not have primary hyperparathyroidism. If these tests are false positive, patients undergo genetic analysis for confirmation and surgery would have been delayed in such patients.

The committee discussed that very low serum magnesium can suppress PTH secretion and also cause resistance to PTH function. This can occur when magnesium is very low and will usually have presented via symptomatic hypomagnesaemia/hypocalcaemia. The committee does not recognise that mild reductions in serum magnesium have material or clinically relevant effects on PTH secretion or function in PHPT.

The committee agreed that all patients with a confirmed diagnosis of primary hyperparathyroidism will need baseline assessment of their symptoms: eGFR/serum creatinine, BMD by DXA scan of lumbar spine, distal radius and hip, and ultrasound of the renal tract to help determine the optimal management pathway.

The committee decided not to recommend phosphate as a separate test, as phosphate is not part of ‘pooled’ biochemistry testing. The committee’s opinion was that phosphate testing was not used as much now as in the past because of the improvement in PTH assays, however this test may be helpful in distinguishing primary hyperparathyroidism from other causes, for example cancer.

The committee decided not to recommend alkaline phosphatase, as it is part of pooled testing and, in any case, was not considered to be of help in establishing the diagnosis of primary hyperparathyroidism.

The committee discussed that there are a number of different genetic tests available and ways they can be performed and this is outside the scope of this guideline; hence they did not make a separate recommendation on these tests. The committee highlighted that these tests are important in assessment of other endocrine conditions such as multiple endocrine neoplasia (MEN).

The committee acknowledged a potential role of bone turnover markers, but due to lack of evidence decided not to recommend, but drafted a research recommendation for the test. The committee discussed that current practice involved using DXA scans to assess fracture risk, but it would be useful to know if bone turnover markers could be used as a surrogate marker for fracture risk. The committee noted that theoretically bone turnover markers could be better than DXA scans in assessing fracture risk.

B.1. Clinical search literature search strategy

Searches were constructed using a PICO framework where population (P) terms were combined with Intervention (I) and in some cases Comparison (C) terms. Outcomes (O) are rarely used in search strategies for interventions as these concepts may not be well described in title, abstract or indexes and therefore difficult to retrieve. Search filters were applied to the search where appropriate.

Table 7. Database date parameters and filters used

Medline (Ovid) search terms

Embase (Ovid) search terms

Cochrane Library (Wiley) search terms

CINAHL (EBSCO) search terms

PsycINFO (ProQuest) search terms

B.2. Health Economics literature search strategy

Health economic evidence was identified by conducting a broad search relating to primary hyperparathyroidism population in NHS Economic Evaluation Database (NHS EED – this ceased to be updated after March 2015) and the Health Technology Assessment database (HTA) with no date restrictions. NHS EED and HTA databases are hosted by the Centre for Research and Dissemination (CRD). Additional searches were run on Medline and Embase for health economics papers published since 2002.

Table 8. Database date parameters and filters used

Medline (Ovid) search terms

Embase (Ovid) search terms

NHS EED and HTA (CRD) search terms

H.1. Excluded clinical studies

Table 9. Studies excluded from the clinical review

H.2. Excluded health economic studies

None.

I.1. Bone turnover markers