Low Prevalence of Exploratory Biochemical Patterns Compatible with Osteomalacia in a Large Cohort of Participants with Low Serum 25-Hydroxyvitamin D Levels

Abstract

The threshold for vitamin D deficiency is mainly based on osteomalacia prevention. Osteomalacia is diagnosed by bone biopsy, but may potentially also be identified biochemically. Our objective was to examine the prevalence of exploratory biochemical patterns compatible with osteomalacia in a population with low vitamin D status, as well as the effect of vitamin D supplementation on the biochemical parameters. The study was cross-sectional followed by a randomized controlled trial (trial registration number: NCT02750293). The participants were recruited from the population-based Tromsø study. Plasma parathyroid hormone, serum calcium, alkaline phosphatase and bone-specific alkaline phosphatase were measured. 423 participants were included in the cross-sectional part and 382 completed the randomized controlled trial (194 given vitamin D (100,000 IU as a loading dose followed by 20,000 IU per week for four months) and 188 given placebo). The mean serum 25-hydroxyvitamin D level in the cohort was 34 nmol/L. None of the participants had a combination of high (above reference range) serum parathyroid hormone, high alkaline phosphatase, and low (below reference range) serum calcium. With the most permissive exploratory biochemical pattern the prevalence was 8.3%. In the intervention study vitamin D significantly reduced plasma parathyroid hormone and alkaline phosphatase and increased serum calcium as compared to placebo. In conclusion, the prevalence of the explored biochemical patterns compatible with ostemalacia was low and strongly dependent on the analytical thresholds applied. In the absence of validation against bone histomorphometry, these findings cannot determine the prevalence of osteomalacia or establish an appropriate serum 25OHD threshold.

Introduction

The main function of vitamin D is to enhance intestinal calcium absorption [1]. Calcium is essential for skeletal health, and in children vitamin D deficiency may lead to rickets and in adults to osteomalacia [1]. There are many indications that vitamin D also has important extra-skeletal effects, but so far, a causal relation has been hard to document. Therefore, when the Institute of Medicine (IOM) in 2011 made their vitamin D intake recommendations for adults, these were mainly based on the relation between serum 25-hydroxyvitamin D (25OHD) level and osteomalacia [2]. This had been studied by Priemel et al. where bone biopsies for histological diagnosis of osteomalacia and blood samples for 25OHD determination were taken post-mortem in 675 participants [3]. IOM’s interpretation of the results was that a serum 25OHD level of 50 nmol/L would protect 97.5% of the cohort from osteomalacia. Subsequently the IOM estimated the vitamin D intake needed to achieve this serum level of 50 nmol/L, and used that in the intake recommendations [2].

Since the study by Priemel et al. there has been no similar publication on osteomalacia and serum 25OHD. One obvious reason is that to diagnose osteomalacia with certainty a bone biopsy is needed [4]. This is an invasive procedure and ethically questionable in a pure research setting. Therefore, criteria for diagnosing osteomalacia biochemically have been suggested [5–8]. This approach is reasonable since there is a distinct biochemical pattern seen together with the development of osteomalacia. Initially, serum calcium is only slightly reduced and usually within the normal range, whereas parathyroid hormone (PTH) and alkaline phosphatase (ALP) levels are more affected and often above the reference range. Then gradually there is a clear decline in serum calcium and possibly also in serum phosphate, with further increases in PTH and ALP [9]. However, as pointed out as early as in 1980 by Nordin, there is no single biochemical variable that is constantly abnormal and also specific for the disease. However, if several variables are taken into account, a strong presumptive diagnosis can be made [10].

Unfortunately, there have been no large studies where levels of PTH, ALP and calcium, or a combination of these, have been systematically evaluated against histological osteomalacia findings. There are, however, two studies where serum PTH, ALP and calcium measured in routine clinical laboratories have been used to estimate prevalence of biochemical osteomalacia with verification against clinical diagnosis in medical records. Thus, in a study from New Zealand including 32,386 subjects, biochemical osteomalacia was found in 0.05% and clinically verified in 0.02% [7]. In a similar study from Scotland including 25,379 subjects, the corresponding frequencies were 0.5%, and 0.1% [8]. The prevalence of biochemical osteomalacia will of course depend on the parameters included and the cut-off values chosen, and in these two studies fairly strict criteria with serum PTH and ALP above the reference range and serum calcium 0.1—0.2 mmol/L below the reference range were used to diagnose biochemical osteomalacia [7, 8].

In Tromsø we have recently performed both a cross-sectional and a four months vitamin D intervention study in 455 participants with a mean baseline serum 25OHD level of 34 nmol/L [11]. In view of the serum 25OHD cut-off of 50 nmol/L applied by the IOM, one would in this cohort expect there to be a considerable number of participants with osteomalacia. Accordingly, depending on criteria chosen, quite a few should test positive for biochemical abnormalities compatible with this diagnosis. The cohort would therefore be suitable for testing different exploratory biochemical patterns compatible with osteomalacia versus serum 25OHD levels, and also to confirm or question the basis of the 25OHD 50 nmol/L cut-off set by the IOM. Furthermore, if the biochemical changes supposedly indicating osteomalacia were caused by vitamin D deficiency, one would expect reversal of these changes in those randomized to vitamin D in the intervention study. If so, this might give support the concept of biochemical osteomalacia.

Materials and Methods

The recruitment of the participants, the design of the study, the results regarding the main endpoint (cardiovascular risk factors), compliance, side effects and causes of drop-out have previously been described in detail [11]. In short, the participants were recruited from the seventh survey of the Tromsø study conducted in 2015/2016 where 20,922 participants had successfully measured serum 25OHD. Those with serum 25OHD below the 10th percentile (< 42 nmol/L) and age < 80 years were invited (n = 1489) to the study. Six hundred and thirty-nine responded and were screened for participation by phone, and 455 came to the baseline visit. Among these, 422 were randomized to vitamin D (100,000 IU cholecalciferol as a loading dose followed by 20,000 IU per week) or identical looking placebo for four months. Effects on bone turn-over markers were secondary endpoints, but the use of these variables to diagnose biochemical osteomalacia had not been prespecified (trial registration number: NCT02750293).

Blood samples were drawn in the fasting state at start and end of the intervention. Serum calcium (reference range 2.15–2.55 mmol/L), serum phosphate (reference range females 0.85–1.50 mmol/L, males 0.75–1.35 mmol/L), estimated glomerular filtration rate (eGFR) (renal impairment < 60 mL/min/1.73 m²), serum aspartate aminotransferase (ASAT) (reference range females 15–35 U/L, males 15—45 U/L), alanine aminotransferase (ALAT) (reference range females 10–45 U/L, males 10—70 U/L), serum ALP (reference range 35- 105 U/L), serum thyroid stimulating hormone (TSH) (reference range 0.2–4.3 mIU/L), and free thyroxine (FT4) (reference range 9–22 pmol/L) were analyzed as previously described [11–14]. PTH was measured in EDTA plasma with the Roche Elecsys PTH intact (Roche Catalog No.: 11972103122) on the Roche Cobas 6000 E601 platform. The antibodies used in this assay are reactive with epitopes in the amino acid regions 26–32 and 37–42. According to the manufacturer's documentation, this assay is standardized against a commercial PTH RIA; recovery of the WHO/NIBSC 95/646 standard has been evaluated, but the assay is not calibrated against this material. In our laboratory, the analytic interassay variation coefficient was < 7%, measured at a plasma PTH level of 4 pmol/L. The laboratory previously used the Immulite assay from Diagnostic Products Corporation with reference range < 50 years 1.1–6.8 pmol/L, > 50 years 1.1–7.5 pmol/L. Before converting to the Roche method, a validation was performed and the two assays gave similar results. The laboratory therefore decided to keep the reference range unchanged. Serum bone-specific alkaline phosphatase (BALP) was measured at Hormone Laboratory, Oslo University Hospital, Aker, Norway using an immunometric chemiluminescence immunoassay (Liaison XL kit from DiaSorin Inc.) (reference range 5.5–25 ug/L). The method was accredited according to NS-EN ISO 15189:2022. Serum 25OHD was measured by an in-house liquid chromatography–tandem mass spectrometry (LC–MS/MS) method that detects both 25OHD₃ and 25OHD₂ and the sum of these presented as 25OHD in the results (reference range 50–150 nmol/L) [15]. The laboratory takes part in the external quality program DEQAS with 75% or more within ± 25% of the target value (performance target set by the DEQAS advisory panel). Participants with serum calcium, ASAT or ALAT above the reference range, hypo- or hyperthyroidism, or eGFR < 60 mL/min/1.73 m², were excluded from the study.

Since our laboratory use reference ranges for plasma PTH slightly higher than that recommended by the producer Roche (1.6–6.9 pmol/L), and since the PTH levels may vary by season and vitamin D status, we included three plasma PTH cut-offs. We also included lower cut-offs for serum ALP and BALP than the upper reference limits used by our local laboratories. Thus, the following exploratory biochemical patterns compatible with osteomalacia were evaluated:

1. Plasma PTH and serum ALP above reference range and serum calcium below reference range

2. Plasma PTH and serum ALP above reference range

3. Plasma PTH above reference range and serum ALP > 85 U/L (arbitrarily chosen)

4. Plasma PTH and serum BALP above reference range and serum calcium below reference range

5. Plasma PTH and serum BALP above reference range

6. Plasma PTH above reference range and serum BALP > 20 ug/L (arbitrarily chosen)

7. Plasma PTH > 6.0 pmol/L (arbitrarily chosen), serum ALP above reference range and serum calcium below reference range

8. Plasma PTH > 6.0 pmol/L and serum ALP above reference range

9. Plasma PTH > 6.0 pmol/L and serum ALP > 85 U/L

10. Plasma PTH > 5.3 pmol/L (upper limit of normality found in healthy subjects with serum 25OHD > 75 nmol/L [16]), serum ALP above reference range and serum calcium below reference range

11. Plasma PTH > 5.3 pmol/L and serum ALP above reference range

12. Plasma PTH > 5.3 pmol/L and serum ALP > 85 U/L

For the evaluation of exploratory biochemical patterns compatible with osteomalacia versus 25OHD levels, a surrogate variable was also created by multiplying the levels of plasma PTH and serum ALP. We did not include serum phosphate among the diagnostic criteria since serum phosphate has a very low sensitivity for diagnosing osteomalacia [8, 17].

Statistical Analyses

Normal distribution was evaluated with visual inspection of histograms and with skewness (between − 1 and 1) and kurtosis (between − 3 and 3). In the intervention study, correlations between baseline and end of study values in the placebo group (tracking), were evaluated with Spearman’s rho. Delta values were calculated as value at end of study minus value at baseline. The delta values in the vitamin D and the placebo group were compared with the independent samples student’s t-test. Prevalence estimates with 95% confidence intervals were calculated with the Clopper-Pearson method.

Normally distributed values are presented as mean ± SD, and non-normally distributed values as median and range (lowest and highest values). All tests were performed double-sided and P < 0.05 was considered statistically significant. IBM SPSS Statistics version 31 was used for the statistical analyses.

Results

Cross-Sectional Study

Among the 455 participants who met at baseline, 453 had complete datasets for the present study. Three participants were excluded because of hypercalcemia, one because of hypothyroidism, one because of hyperthyroidism, three because of eGFR < 60 mL/min/1.73 m², and an additional 22 participants because of elevated liver transaminases leaving 423 participants (191 females, 232 males) for the cross-sectional analyses. Their baseline characteristics in relation to serum 25OHD are shown in Table 1. The distribution of serum 25OHD is shown in Fig. 1.

Table 1.

Baseline characteristics in the cross-sectional study in relation to serum 25OHD levels

	All participants N = 423	Serum 25OHD
		< 50 nmol/L N = 375	< 30 nmol/L N = 174	< 25 nmol/L N = 98
Males/females	232/191	202/173	99/75	56/42
Age (years)	52 (40, 79)	52 (40, 79)	51 (40, 76)	50 (40, 76)
BMI (kg/m2)	27.9 ± 5.1	28.1 ± 5.1	28.1 ± 5.3	28.2 ± 5.3
eGFR (mL/min/1.73 m2)	96 (62, 127)	96 (62, 127)	97 (65, 127)	97 (65, 127)
Serum calcium (mmol/L)	2.27 ± 0.07	2.27 ± 0.07	2.27 ± 0.07	2.27 ± 0.07
Serum phosphate (mmol/L)	1.00 ± 0.16	1.00 ± 0.16	0.99 ± 0.15	0.97 ± 0.15
Plasma PTH (pmol/L)	6.4 (2.7, 15.1)	6.5 (2.7, 15.1)	6.8 (3.5, 15.1)	7.0 (3.5, 15.1)
Serum ALP (U/L)	66 ± 17	66 ± 17	66 ± 16	66 ± 15
Serum BALP (ug/L)	10.7 ± 3.9	10.8 ± 4.1	10.7 ± 4.0	10.9 ± 3.6
Serum 25OHD (nmol/L)	33.9 ± 12.7	30.6 ± 8.7	22.9 ± 4.8	19.6 ± 3.6

Data are presented as mean ± SD or median (minimum, maximum)

BMI, body mass index; eGFR, estimated glomerular filtration rate; PTH, parathyroid hormone; ALP, alkaline phosphatase; BALP, bone-specific alkaline phosphatase; 25OHD, 25-hydroxyvitamin D

Fig. 1.

Distribution of serum 25-hydoxyvitamin D in the 423 participants in the cross-sectional study

There were several inter-correlations between the exploratory biochemical osteomalacia variables. Of note, there were for plasma PTH positive correlations with serum ALP and BALP, and negative with serum 25OHD. On the other hand, there was for serum 25OHD no significant correlations with serum ALP and BALP (Table 2).

Table 2.

Spearman’s correlation coefficient rho in the 423 participants at baseline

	Serum calcium	Plasma PTH	Serum ALP	Serum BALP	Serum 25OHD
Serum calcium		− 0.142*	0.136*	0.138*	0.047
Plasma PTH	− 0.142*		0.161**	0.190**	− 0.191**
Serum ALP	0.136*	0.161**		0.788**	− 0.039
Serum BALP	0.138*	0.190**	0.788**		− 0.051
Serum 25OHD	0.047	− 0.191**	− 0.039	− 0.051

ALP, alkaline phosphatase; BALP, bone-specific alkaline phosphatase; PTH, parathyroid hormone; 25OHD, 25-hydroxyvitamin D

*P < 0.01; **P < 0.001

Among the 423 participants, 134 (31.7%) had plasma PTH above the reference range (> 6.8 or 7.5 pmol/L), 237 participants (56.0%) had plasma PTH > 6.0 pmol/L, and 309 participants (73.0%) had PTH > 5.3 pmol/L. Fifteen participants (3.5%) had serum calcium < 2.15 mmol/L and 11 participants (2.6%) had serum ALP > 105 U/L. If changing the ALP cut-off from 105 to 85 U/L, 45 participants (10.6%) would have an elevated level. Two participants had serum BALP above the reference range (25 ug/L), and nine participants had serum BALP > 20 ug/L.

The number of participants found with an exploratory biochemical pattern compatible with osteomalacia depended on the cut-offs applied (Table 3). Thus, not a single participant had a combination of high plasma PTH, high serum ALP (or serum BALP) and low serum calcium. Only four (0.9%) had a combination of high plasma PTH and high serum ALP, and none a combination of high plasma PTH and serum BALP. Eighteen participants (4.3%) had a combination of high plasma PTH and serum ALP > 85 U/L. With the most liberal criteria (plasma PTH > 5.3 pmol/L and serum ALP > 85 U/L) 35 participants (8.3%) had an exploratory biochemical pattern compatible with osteomalacia.

Table 3.

Prevalence of participants with exploratory biochemical patterns compatible with osteomalacia using different criteria in relation to serum 25OHD level

	All participants N = 423	Serum 25OHD < 50 nmol/L N = 375	Serum 25OHD < 30 nmol/L N = 174	Serum 25OHD < 25 nmol/L N = 98
Criteria 1	0.000 (0.000–0.009)	0.000 (0.000–0.010)	0.000 (0.000–0.021)	0.000 (0.000–0.037)
Criteria 2	0.009 (0.003–0.024)	0.011 (0.003–0.027)	0.006 (0.000–0.032)	0.010 (0.000–0.056)
Criteria 3	0.043 (0.025–0.066)	0.048 (0.029–0.075)	0.063 (0.032–0.110)	0.082 (0.036–0.155)
Criteria 4	0.000 (0.000–0.009)	0.000 (0.000–0.010)	0.000 (0.000–0.021)	0.000 (0.000–0.037)
Criteria 5	0.000 (0.000–0.009)	0.000 (0.000–0.010)	0.000 (0.000–0.021)	0.000 (0.000–0.037)
Criteria 6	0.007 (0.001–0.021)	0.008 (0.002–0.023)	0.006 (0.000–0.032)	0.010 (0.000–0.056)
Criteria 7	0.000 (0.000–0.009)	0.000 (0.000–0.010)	0.000 (0.000–0.021)	0.000 (0.000–0.037)
Criteria 8	0.019 (0.008–0.037)	0.021 (0.009–0.042)	0.017 (0.004–0.050)	0.010 (0.000–0.056)
Criteria 9	0.076 (0.052–0.105)	0.085 (0.059–0.118)	0.098 (0.058–0.152)	0.112 (0.057–0.192)
Criteria 10	0.000 (0.000–0.009)	0.000 (0.000–0.010)	0.000 (0.000–0.021)	0.000 (0.000–0.036)
Criteria 11	0.021 (0.010–0.040)	0.021 (0.009–0.042)	0.017 (0.004–0.050)	0.010 (0.000–0.056)
Criteria 12	0.083 (0.058–0.113)	0.091 (0.064–0.124)	0.098 (0.058–0.152)	0.112 (0.057–0.192)

Data are presented as prevalence (95% confidence interval)

PTH, parathyroid hormone; ALP, alkaline phosphatase; BALP, bone-specific alkaline phosphatase; 25OHD, 25-hydroxyvitamin D

Criteria 1: Plasma PTH and serum ALP above reference range and serum calcium below reference range

Criteria 2: Plasma PTH and serum ALP above reference range

Criteria 3: Plasma PTH above reference range and serum ALP > 85 U/L

Criteria 4: Plasma PTH and serum BALP above reference range and serum calcium below reference range

Criteria 5: Plasma PTH and serum BALP above reference range

Criteria 6: Plasma PTH above reference range and serum BALP > 20 ug/L

Criteria 7: Plasma PTH > 6.0 pmol/L, serum ALP above reference range and serum calcium below reference range

Criteria 8: Plasma PTH > 6.0 pmol/L and serum ALP above reference range

Criteria 9: Plasma PTH > 6.0 pmol/L and serum ALP > 85 U/L

Criteria 10: Plasma PTH > 5.3 pmol/L, serum ALP above reference range and serum calcium below reference range

Criteria 11: Plasma PTH > 5.3 pmol/L and serum ALP above reference range

Criteria 12: Plasma PTH > 5.3 pmol/L and serum ALP > 85 U/L

The “surrogate” biochemical osteomalacia variable (plasma PTH x serum ALP) was significantly associated with serum 25OHD (rho − 0.158, P = 0.001) and shown graphically in Fig. 2.

Fig. 2.

Scatter plot of serum 25-hydroxyvitamin D versus the exploratory biochemical osteomalacia “surrogate” variable plasma parathyroid hormone (PTH) x serum alkaline phosphatase (ALP) in the 423 participants in the cross-sectional study

Intervention Study

Of the 423 participants included at baseline, 399 were randomized to vitamin D or placebo, and 382 completed the study (194 in the vitamin D group and 188 in the placebo group). The baseline, end of study, and delta values of these 382 participants are shown in Table 4. As compared to the placebo group, there was in the vitamin D group at the end of the intervention a significant decrease in plasma PTH, serum ALP and BALP, and a significant increase in serum calcium.

Table 4.

Participants’ characteristics and baseline and delta values in the intervention study

	Baseline vitamin D N = 194	Baseline placebo N = 188	End of study delta values vitamin D N = 194	End of study delta values placebo N = 188	Mean difference (95% CI) delta values
Males/females	108/86	100/88
Age (years)	49 (40, 79)	51 (40, 76)
BMI (kg/m2)	27.7 ± 5.0	27.8 ± 4.7
eGFR (mL/min/1.73 m2)	99 (65, 127)	96 (62, 121)
Serum calcium (mmol/L)	2.27 ± 0.07	2.27 ± 0.07	0.01 ± 0.07**	− 0.01 ± 0.06	0.02 (0.01–0.03)
Serum phosphate (mmol/L)	1.00 ± 0.16	1.00 ± 0.16	0.03 ± 0.15	0.02 ± 0.14	0.01 (− 0.02 to 0.03)
Plasma PTH (pmol/L)	6.2 (2.7, 15.1)	6.5 (3.1, 12.9)	− 0.8 ± 1.4***	0.5 ± 1.5	− 1.3 (− 1.6 to − 1.0)
Serum ALP (U/L)	65 ± 17	65 ± 17	0.6 ± 7.5*	2.3 ± 7.8	− 1.6 (− 3.1 to − 0.1)
Serum BALP (ug/L)	10.9 ± 4.1	10.6 ± 3.6	− 0.3 ± 1.8**	0.2 ± 1.9	− 0.5 (− 0.9 to − 0.2)
Serum 25OHD (nmol/L)	32.5 ± 10.8	35.3 ± 13.5	56.5 ± 21.7***	− 4.5 ± 12.8	61.0 (57.4–64.6)

Data are presented as mean ± SD or median (minimum, maximum)

BMI, body mass index; CI, confidence interval; eGFR, estimated glomerular filtration rate; PTH, parathyroid hormone; ALP, alkaline phosphatase; BALP, bone-specific alkaline phosphatase; 25OHD, 25-hydroxyvitamin D

*P < 0.05; **P < 0.01; ***P < 0.001 (versus placebo group, independent samples student’s t-test). Delta values are value at end of study minus value at baseline

In the placebo group the variables showed a high degree of tracking (correlation between baseline and end of study value), the Spearman’s rho for plasma PTH, serum ALP, BALP, calcium and 25OHD being 0.69, 0.92, 0.88, 0.64, and 0.40, respectively (P < 0.001).

Discussion

In the present study we have found what appears to be a low prevalence of exploratory biochemical patterns compatible with osteomalacia in a cohort of participants with low serum 25OHD levels.

This was surprising in view of the study by Priemel et al. where bone biopsies for histological examination were taken from 675 participants [3] and more than 25% were diagnosed with osteomalacia based on percentage osteoid volume. Our cohort was similar to that of Priemel et al. regarding male/female ratio (54.6% and 59.4%, respectively), but slightly younger (mean age 52 and 63 years, respectively). However, according to Priemel et al. the prevalence of osteomalacia was not related to age [3]. On the other hand, the difference in vitamin D status was obviously important as our cohort had a substantially higher mean serum 25OHD level (34 nmol/L versus ~ 20 nmol/L). We therefore looked closely at the scatterplot with percentage osteoid volume versus serum 25OHD in the Priemel et al. publication, and selected a sub-group with a better vitamin D status than our cohort. Thus, by counting the dots on the plot, we found approximately 300 participants with serum 25OHD in the range 25–125 nmol/L, and among these, approximately 43 participants (14.3%) were diagnosed with osteomalacia [3]. An illustrative estimate would therefore be that approximately 60 participants in our cohort should have osteomalacia. However, even our most “liberal” biochemical pattern did not identify more than 35.

The Priemel et al. study has been criticized for several reasons, including their histological criteria for osteomalacia, lack of clinical data, post-mortem blood sampling, and the use of a non-standardized 25OHD assay [18]. Thus, there could be several explanations for the apparent discrepancy between their and our findings. Firstly, the 675 participants studied by Priemel et al. had surprisingly low serum 25OHD levels even for a German population [19]. One might suspect that this could be due to their blood samples being drawn post-mortem or to the use of a 25OHD immunoassay that might not have been as accurate as modern LC–MS/MS methods. However, if the 25OHD levels were falsely low, “correct” (higher) values would have resulted in an even higher expectancy of osteomalacia in our cohort. A more likely explanation could be that the histological osteomalacia criteria used by Priemel et al. were set too low. They used a cut-off for osteomalacia as percentage osteoid volume (osteoid volume / bone volume (OV/BV)) > 2% [3]. Unfortunately, there are no large studies on OV/BV in healthy participants, but in the few available on participants without symptoms of or known skeletal disease, or evidence of organ disease by autopsy, the mean OV/BV has been close to 3% with a SD of ~ 1.5% [20–24]. Accordingly, the upper “normal” limit should be close to 5%, which has also been suggested by Bahn and Minisola [4, 9]. It should be noted that the cut-off suggested by Bahn was to diagnose the early stages of osteomalacia (“pre-osteomalacia”) before irreversible bone loss had occurred [4]. Furthermore, both Bahn and Minisola suggested that the diagnosis should not only rely on one bone parameter, but on the combination of two, for instance OV/BV and osteoid thickness [4, 9]. If so, the threshold for osteomalacia diagnosis would be even higher. Thus, if applying an osteoid volume of 5% as cut-off for osteomalacia on the Priemel et al. data, and looking at those with serum 25OHD in the range 25–125 nmol/L, then apparently only seven out of the approximately 300 participants (2.3%) in their study would be diagnosed with osteomalacia [3]. Accordingly, there should be approximately ten participants in our cohort with osteomalacia.

However, using the standard criteria for biochemical osteomalacia with elevated serum PTH and ALP and reduced serum calcium [7, 8], not a single participant was identified in our cohort. This may indicate that these standard criteria for biochemical osteomalacia are set too strict.

It is well known that in patients with osteomalacia the plasma PTH and serum ALP levels are high and the serum calcium and phosphate levels are low. This was shown in a study by Basha et al. where 15 patients with osteomalacia due to vitamin D depletion and diagnosed because of clinical suspicion, all had elevated PTH and ALP levels [21]. However, even in patients with overt disease, not everyone has biochemical levels outside the reference range. Thus, in a study by Peach et al. including 50 patients with histologically proven osteomalacia (defined as osteoid volume > 10% and calcification fronts < 60%), serum ALP was within the reference range in 14%, serum calcium in 38% and serum phosphate in 96% of the participants [17].

Due to this low sensitivity of serum calcium and phosphate to detect osteomalacia, and since it is important to diagnose osteomalacia in its early stages before irreversible changes occur [4], it might seem reasonable not to include calcium and phosphate in the potential biochemical osteomalacia criteria. As an indicator of calcium status, plasma PTH is much more sensitive than serum calcium as even small decreases in serum calcium within the reference range cause substantial and compensatory increases in plasma PTH [25]. In addition to an increase in the plasma PTH level, indicating calcium deficiency, there should also be a measure indicating that in spite of a compensatory PTH increase, the calcium deficiency had resulted in increased osteoblast activity. For this, serum ALP measurement is widely available and has been the traditional choice. Accordingly, we suggest that plasma PTH and serum ALP are candidate markers for biochemical osteomalacia diagnosis requiring validation against histomorphometry.

In addition to serum ALP we also measured serum BALP and found approximately similar results, which might be due to the exclusion of participants with potential renal and hepatic conditions. In an unselected population, the use of serum BALP might therefore be preferable.

And finally, another explanation for the apparent discrepancy between the Priemel study and ours, could be differences in calcium intake. In the pathogenesis of nutritional osteomalacia there is (as for rickets) an interaction between calcium intake and vitamin D, where a high calcium intake is protective [26, 27]. However, the calcium intake in Northern Norway and Germany seems to be fairly similar although differences between the two cohorts cannot be excluded [28–30].

When setting the biochemical criteria, one must not only consider sensitivity, but also specificity. If excluding serum calcium from the criteria (as we suggest), that would of course increase the sensitivity at the cost of specificity, as would lowering the PTH and ALP cut-offs.

In the interventional part of our study, we found as expected in the vitamin D group compared to the placebo group, a significant reduction in plasma PTH, serum ALP and BALP, and an increase in serum calcium. In subjects with osteomalacia, biochemical variables typically begin to normalize within a few weeks to three months after starting vitamin D treatment [31]. Therefore, the reduction in plasma PTH and ALP by vitamin D found by us, as well as their high degree of tracking, could strengthen their role as candidate markers in biochemical osteomalacia diagnosis. However, although the effects of vitamin D on these markers were statistically significant, the magnitude was small and could well be due to a physiological effect of vitamin D and not be osteomalacia-reversal specific.

We created a “surrogate” biochemical parameter by multiplying plasma PTH with serum ALP in the hope that this parameter would better identify participants with potential osteomalacia. If so, when plotted against serum 25OHD, one might detect a deflection point for significant vitamin D deficiency. This PTH*ALP parameter certainly correlated with serum 25OHD but was not even close to indicate a 25OHD cut-off value. This could be because our cohort, in spite of low serum 25OHD values, was not truly vitamin D deficient, or was protected by a high calcium intake, or that biochemical osteomalacia might not properly identify true osteomalacia.

This does not mean that the concept of biochemical osteomalacia should be discarded, but before it can be useful it has to be evaluated against the gold standard, histological examination of bone biopsies. Such a bone biopsy study is needed not only for establishing biochemical criteria, but also for re-evaluating the relation between histologically verified osteomalacia and serum 25OHD levels. This is important since two recent publications have found no significant correlation between serum 25OHD levels and indices of bone mineralization, at least in participants without severe vitamin D deficiency [32, 33]. Such a study would also be valuable in the ongoing discussion on which vitamin D metabolite(s) is the best marker for vitamin D sufficiency [34]. The optimal measure would be the intra-cellular concentration of the active hormone 1,25-dihydroxyvitamin D, which at present is unavailable and could also differ between tissues. Other candidates could be cholecalciferol itself or metabolite ratios like the one between 25OHD and 24,25dihydroxyvitamin D [35, 36]. However, this question can hardly be settled without comparison with histologically proven osteomalacia.

If criteria for biochemical osteomalacia can be established with acceptable sensitivity and specificity, it could be useful not only for finding 25OHD cut-offs for vitamin D sufficiency in various calcium-intake cohorts, but also for evaluating a serum 25OHD measurement in a clinical setting. It could also be used as inclusion and safety criterium to avoid ethical problems in long-lasting vitamin D trials.

It should be emphasized that our study has several limitations, the main one being that our study did not include bone biopsies. When we compared with the Priemel et al. study we counted dots on a published scatter plot which obviously is not an exact method and the numbers found must be taken as approximations and illustrations only [3]. The present analyses were not predefined and must therefore be considered as exploratory, in particular our choice of biochemical patterns compatible with osteomalacia. Also, calcium intake was not recorded, and we measured total and not the preferred serum ionized calcium. We excluded participants with hypercalcemia, renal impairment, thyroid dysfunction and elevated transaminases, participants that might well be at increased risk of osteomalacia. This may have reduced the prevalence of participants with exploratory biochemical patterns compatible with osteomalacia, and our results cannot be extrapolated to more frail populations. There are also some discussions as to whether the current immunoassays for PTH are precise enough [37]. And finally, although our serum 25OHD measurements were performed by LC–MS/MS in a laboratory participating in the DEQAS program, samples were not reanalyzed with a method traceable to the Centre for Disease Control Vitamin D Standardization-Certification Program. On the other side, we had a large cohort of participants with low serum 25OHD levels, and our potential markers of biochemical osteomalacia could be tested for vitamin D responsiveness.

In conclusion, we have in a large cohort with low serum 25OHD levels found an apparently low prevalence of participants with exploratory biochemical patterns compatible with osteomalacia. However, since we did not have bone biopsies in our study, our findings should be interpreted with great caution. To draw firm conclusions, new and large studies with bone biopsies where histological diagnosis of osteomalacia is related to serum 25OHD levels and other markers of calcium metabolism are highly needed.

Funding

Open access funding provided by UiT The Arctic University of Norway (incl University Hospital of North Norway). The present study was supported by grants from the North Norway Regional Health Authorities (grant number 6856/SFP1029-12) and UiT The Arctic University of Norway.

Data Availability

Restrictions apply to the availability of data generated during this study to preserve patient confidentiality. The corresponding author will on request detail conditions under which access to some data may be provided.

Declarations

Conflict of interest

None.