The influence of sampling time on indirect reference limits, decision limits, and the estimation of biological variation of random plasma glucose concentrations

Introduction

The biological variation in plasma glucose levels results from the complex interaction of genetically anchored metabolic processes that are subject to strict hormone-controlled regulation. From the evolutionary point of view, humans were enabled to sleep through the night as a prerequisite for brain consolidation [1]. Physiological stimuli are set by the environment which function as synchronizer or so-called Zeitgeber (literally “time giver”) but can be modified by pathophysiological factors [2]. As a result, plasma glucose levels exhibit pronounced daytime-dependent fluctuations that are not solely due to nutritional intake [3], [4].

The time-dependence of glucose levels affects laboratory diagnostics at various aspects. The most obvious is the relation between blood sampling time and time of food intake, i.e., fasting and postprandial status as well as randomly collected samples [5]. Clinical decision limits exist for all the respective prandial states. In addition to this extrinsic factor, there are intrinsic regulatory mechanisms that have developed evolutionarily and determine a centrally anchored diurnal rhythm of glucose metabolism independent of food intake [6]. These intrinsic mechanisms contribute to a certain extent systematically and predictably to the biological variability of glucose levels in a population.

Therefore, we have processed large-scale laboratory data to characterize the predictable, i.e., non-random, variability of plasma glucose levels and to estimate their influence on reference intervals, decision limits and biological variability.

Materials and methods

Results

Hourly distribution of plasma glucose levels

Table 1 summarizes the underlying samples sizes and their distribution in the respective time phases. More than two third of all samples are collected between 10:00 h in the morning and 17:59 h in the evening and thus outside the time frame recommend for the collection of fasting plasma glucose which is 07:00–09:00 h [5].

Table 1:

Distribution of random glucose samples stratified by study site, gender, age, and distinct time intervals.

Site 1 (UKSH)
Gender, n	Age classes, years (n)	Daytime	n (%)	6.1–6.9 mmol/L	7.0–8.0 mmol/L	≥11.1 mmol/L
				n (%)a	n (%)b	n (%)b
Male 263,480	18–39 (36,808)	18:00–05:59	8,529 (23.2)	n/ac	n/a	171 (2.0)
		06:00–09:59	3,786 (10.3)	532 (14.0)	224 (5.9)	57 (1.5)
		10:00–17:59	24,493 (66.5)	n/a	n/a	445 (1.8)
	40–59 (76,479)	18:00–05:59	11,059 (14.5)	n/a	n/a	674 (6.1)
		06:00–09:59	10,006 (13.1)	1,669 (16.7)	n/a	511 (5.1)
		10:00–17:59	55,414 (72.4)	n/a	n/a	2,606 (4.7)
	≥60 (150,193)	18:00–05:59	21,819 (14.5)	n/a	n/a	1,994 (9.1)
		06:00–09:59	20,220 (13.5)	3,589 (17.7)	n/a	1,640 (8.1)
		10:00–17:59	108,154 (72.0)	n/a	n/a	7,148 (6.6)
Female 250,202	18–39 (44,324)	18:00–05:59	9,672 (21.8)	n/a	n/a	128 (1.3)
		06:00–09:59	4,348 (9.8)	380 (8.7)	141 (3.2)	52 (1.2)
		10:00–17:59	30,304 (68.4)	n/a	n/a	359 (1.2)
	40–59 (67,869)	18:00–05:59	8,712 (12.8)	n/a	n/a	295 (3.4)
		06:00–09:59	7,710 (11.4)	1,111 (14.4)	n/a	240 (3.1)
		10:00–17:59	51,447 (75.8)	n/a	n/a	1,381 (2.7)
	≥60 (138,009)	18:00–05:59	22,189 (16.1)	n/a	n/a	1,763 (7.9)
		06:00–09:59	15,400 (11.1)	2,540 (16.5)	n/a	1,171 (7.6)
		10:00–17:59	100,420 (72.8)	n/a	n/a	5,094 (5.1)
Site 2 (UKK)
Male 102,641	18–39 (23,827)	18:00–05:59	5,392 (22.6)	n/ac	n/a	66 (1.2)
		06:00–09:59	2,781 (11.7)	251 (9.0)	97 (3.5)	16 (0.6)
		10:00–17:59	15,654 (65.7)	n/a	n/a	112 (0.7)
	40–59 (29,416)	18:00–05:59	5,519 (18.8)	n/a	n/a	240 (4.3)
		06:00–09:59	3,074 (10.4)	396 (12.9)	n/a	80 (2.6)
		10:00–17:59	20,823 (70.8)	n/a	n/a	655 (3.1)
	≥60 (49,398)	18:00–05:59	8,007 (16.2)	n/a	n/a	788 (9.8)
		06:00–09:59	4,207 (8.5)	697 (16.6)	n/a	291 (6.9)
		10:00–17:59	37,184 (75.3)	n/a	n/a	2,494 (6.7)
Female 101,360	18–39 (30,660)	18:00–05:59	5,988 (19.5)	n/a	n/a	55 (0.9)
		06:00–09:59	4,183 (13.6)	213 (5.1)	60 (1.4)	21 (0.5)
		10:00–17:59	20,489 (66.8)	n/a	n/a	97 (0.5)
	40–59 (28,452)	18:00–05:59	4,151 (14.6)	n/a	n/a	109 (2.6)
		06:00–09:59	2,986 (10.5)	289 (9.7)	n/a	35 (1.2)
		10:00–17:59	21,315 (74.9)	n/a	n/a	324 (1.5)
	≥60 (42,248)	18:00–05:59	6,721 (15.9)	n/a	n/a	516 (7.7)
		06:00–09:59	3,278 (7.8)	484 (14.8)	n/a	157 (4.8)
		10:00–17:59	32,249 (76.3)	n/a	n/a	1,471 (4.6)

1. Prevalence of patients that violate clinical decision limits: ^aDecision limits refer to current recommendation for the diagnosis of type-2 diabetes [23], [24] and ^breported in the literature for MODY in combination with additional evidence from other criteria [25]. ^cNot applicable for the respective age group or daytime.

Box plots in Figure 1 illustrate the hourly distribution of glucose levels of site 1 data set stratified by gender and age groups. Comparable pattern can be obtained from data of site 2 (data not shown). All six subgroups displayed in Figure 1 demonstrated similar diurnal fluctuations in glucose levels. The highest glucose levels were recorded consistently across all subgroups during the early morning hours between 02:00 and 06:00 h. As expected, glucose levels differed between age classes and gender with increasing plasma concentrations in the elderly and higher values in males (Table 2), respectively.

Figure 1:

Box plots of hourly glucose levels at site 1 grouped by gender and age.

Males (left) at age classes ≥60 (A), 40–59 (B) and 18–39 (C), females (right) at age classes ≥60 (D), 40–59 (E) and 18–39 (F) years of age, respectively. Similar distributions resulted at site 2 (data not shown). Horizontal colored lines represent clinical decision limits for fasting and random glucose levels, respectively. Glucose values between 5.6 (blue) and 6.9 mmol/L (green) are suspect for impaired fasting glucose [23], [24]. Mild fasting hyperglycemia (glucose <8 mmol/L) (orange) in young patients may indicate the presence of maturity onset diabetes of the young (MODY) if there is additional evidence from other criteria [25]. Glucose values above 11.1 mmol/L (red) are suspect for diabetes mellitus.

Table 2:

Rhythmometrics of random glucose levels estimated from laboratory data of two university hospitals.

Site	Gender	Age class, years	Percentiles
			25%				50%				75%
			t 0	M	A	±A/M%	t 0	M	A	±A/M%	t 0	M	A	±A/M%
UKSH	Male	18–39	02:52	4.97	0.40	±8.0	02:14	5.49	0.46	±8.4	01:38	6.12	0.54	±8.8
		40–59	02:52	5.20	0.40	±7.7	02:14	5.81	0.46	±7.9	01:38	6.67	0.54	±8.1
		≥60	02:52	5.40	0.40	±7.4	02:14	6.14	0.46	±7.5	01:38	7.24	0.54	±7.5
	Female	18–39	02:26	4.74	0.34	±7.2	01:41	5.20	0.32	±6.2	01:01	5.79	0.38	±6.6
		40–59	02:51	5.04	0.45	±8.9	02:06	5.61	0.47	±8.4	01:20	6.39	0.55	±8.6
		≥60	02:05	5.40	0.49	±9.1	01:32	6.14	0.57	±9.3	01:07	7.17	0.66	±9.2
UKK	Male	18–39	02:00	4.85	0.31	±6.4	01:14	5.34	0.42	±7.9	00:29	5.94	0.54	±9.1
		40–59	02:00	5.05	0.31	±6.1	01:14	5.63	0.42	±7.5	00:29	6.40	0.54	±8.4
		≥60	02:00	5.27	0.31	±5.9	01:14	5.96	0.42	±7.0	00:29	7.00	0.54	±7.7
	Female	18–39	01:03	4.69	0.32	±6.8	00:16	5.14	0.39	±7.6	23:25	5.68	0.51	±9.0
		40–59	01:03	4.92	0.32	±6.5	00:16	5.42	0.39	±7.2	23:25	6.08	0.51	±8.4
		≥60	01:03	5.22	0.32	±6.1	00:16	5.86	0.39	±6.7	23:25	6.83	0.51	±7.5

1. Acrophases (t ₀) indicate the time (hh:mm) at highest (peak) glucose concentrations provided from two distinct medical university laboratory sites in Germany. MESOR (M) and amplitude (A) of glucose concentrations in mmol/L at the interquartile range (25,50 and 75%) as derived from quantile regression. A/M% indicates the relative deviation of the amplitude from the MESOR and thereby serves as a measure of the anticipated non-random portion of biological variability.

Quantile-based cosinor analyses of diurnal glucose variations

The hourly quantiles (2.5, 25, 50, 75% and 97.5 percentile) were modeled applying QR analysis regressed on two predictor variables, t (hour) as cosinor-function and age classes. Figure 2 depicts the resulting wave-shaped regression lines for site 1 over a 24 h period for each respective quantile. In contrast to data from site 1 no statistically significant cosine function could be determined at the 2.5% percentile in all subgroups of site 2, so that no time-dependent variability of glucose values can be detected in all age groups (data not shown).

Figure 2:

Estimated diurnal rhythm of glucose values during 24 h via quantile regression method for males (left) of age classes ≥60 (A), 40–59 (B), and 18–39 years (C), and females (right) of age classes ≥60 (D) 40–59 (E), and 18–39 years (F) for site 1.

Lines indicate from the bottom to the top the 2.5, 25, 50, 75, and 97.5% percentiles, respectively.

At both sites and in all age groups, acrophases of females tended to occur about 30–60 min earlier as compared to males in the inner interquartile range. According to this model the predictable biological variation inherent in the data sets correspond to the amplitude to MESOR ratio (A/M). Table 2 summarizes the time adjusted quantiles grouped by study site, gender and age classes for median and the upper and lower border of the interquartile range. A/M-ratio steadily increased independent from site, gender and age class from 25 to 75% percentile. Even though a positive correlation between glucose levels and age also exists within each quantile there was no systematic effect detectable on the A/M-Ratio. The overall A/M-ratio achieved 7.7% (range 5.9–9.3%). The highest portion of variability is apparent between 06:00 and 10:00 h. According to the calculated QR model applied glucose levels decrease on average by about 4% in this period, i.e., on average 1% per hour.

Daytime-specific reference limits

According to the dynamics described above, daytime-specific continuous reference intervals were determined. Figure 3 illustrates the resulting upper and lower RL estimated for both gender between 18 and 59 years at both sites, separately. Due to small sample sizes RLs do not cover the whole night phase. Assuming, that published RLs for glucose refer to fasting states the smallest distance between the lower and upper reference limit appear in the morning. The distance between the two limits gradually increases between the afternoon and evening hours.

Figure 3:

Daytime-dependent continuous reference limits (RLs) estimated by the truncated maximum likelihood (TML) approach.Lines indicate the respective lower RL and upper RL of patients aged between 18 and 59 years.

The crosses between upper RL and lower RL represent the corresponding medians. Right: males (blue) at site 1 (A) and site 2 (B); left: females (red) at site 1 (C) and site 2 (D). Corresponding points display point estimations for lower RL, upper RL and median.

Table 3 provides an overview of the estimated reference limits and medians that would result from the TML approach for day (09:00–17:59 h) and nighttime (18:00–04:59 h). Comparing the distribution of glucose values within each hour of Figure 1 with those in Table 3 and Figure 3, the effect of the RLE procedure is evident. This algorithm is designed to extract from a mixed population the “central” part of the distribution that distinguishes non-diseased from diseased individuals. Thus, significantly tighter intervals result as compared to Figure 1. The differences of the medians and upper RL in Table 3 exceeded the respective permissible analytical uncertainty and would therefore have to be considered relevant. In contrast, the lower RL did not or only to a lesser extent differ between day and nighttime.

Table 3:

Daytime and night specific reference limits for males and females derived from random glucose levels at site 1 and 2.

Site	Gender	Glucose, mmol/L
		Day 9:00–16:59				Night 18:00–4:59				Difference (night–day)
		n	Lower RL	Upper RL	Median	n	Lower RL	Upper RL	Median	Δ-Lower RL	Δ-Upper RL	Δ-Median
UKK	Male	39,639	4.00b	6.43c	5.07	10,642	4.11	7.35	5.51	0.11	0.92a	0.44a
	Female	46,674	3.85	6.15	4.88	9,873	3.98	6.64	5.28	0.13	0.49a	0.40a
UKSH	Male	83,511	3.93	6.48	5.20	18,913	4.13	7.29	5.50	0.20a	0.81a	0.30a
	Female	84,837	3.71	6.29	5.00	17,766	3.91	6.64	5.28	0.20a	0.35a	0.28a

1. ^aIndicates that Δ is higher than pU _a . Two explanatory examples are given in the following: ^b pU(4.00)=0.18 mmol/L, this results in a permissible Δ-lower RL for 4.00 mmol/L measured at day. Thus, the estimated value of 4.11 mmol/L at night is equivalent to 4.00 mmol/L, ^c pU(6.43)=0.27 mmol/L is the permissible Δ-upper RL for 6.43 mmol/L measured at day. Thus, the estimated value of 7.35 mmol/L at night is not equivalent to 6.43 mmol/L.

Discussion

Clinical assessment of plasma glucose concentration is based on clinical decision limits. The availability of population-specific reference limits nevertheless has a justification. In particular, for clinical decision limits, it should be known how they relate to the distribution of an analyte in a population. Age, sex, and, as described here, time-dependent variations are, in our opinion, basic information that merit consideration in the design of studies to determine or adjust clinical decision limits. Patient-based real time quality control is another potential application where time-based population-specific reference limits may play a role. In these contexts, our study contributes, yet not available information on the dynamics and magnitude of time-dependent variations of plasma glucose.

The term chronodesm describes time-bound limits within which the measured values of an analyte are physiologically located over a defined period [14]. In clinical practice, this approach is rarely applied to laboratory parameters with known circadian rhythm. So-called monodesms are much more common, albeit this term is not commonly accepted in the context of RL and clinical decision limits. Being a derivate of the term chronodesm, monodesms describe the physiological range or reference interval at a defined point in time or, if applicable, over a circumscribed phase that is located within a period of the principal rhythm. In clinical practice, therefore, monodesms actually reflect a pre-analytical requirement for the time of blood collection.

In the analyses presented here, the diurnal dynamics of random glucose levels were addressed under different aspects. In contrast to time series obtained from single persons large-scale “real-world” laboratory data have been investigated.

The term real-world big data study was coined to distinguish such approaches from conventional controlled trials in well-characterized cohorts [15]. Its aim is to counter the criticism of artificial settings in controlled trials by generating findings that better reflect the situation in everyday clinical practice, in the sense of “real world evidence” [16].

Studies on the time-dependent variation of plasma glucose are mostly restricted to either healthy young volunteers or patients suffering from disorders in glucose metabolism. Studies that apply time series approaches are usually, however, limited to smaller sample sizes. The latter focus mainly on the estimation of day-to-day variability and analyze fasting glucose levels obtained under strict pre-analytical conditions [17]. We, therefore, assume that metrics provided from real-world laboratory data will not fit to any type of variability generated by those studies.

Data included in our analyses represented only first values. In addition to the exclusion of any bias due to autocorrelation or interventions such as medication or hemodialysis, this type of data refinement is decisive for the conclusions drawn. The extent of individual influences on the components of biological variability inherent in the data sets shall be reduced to those that can be considered as population specific. We further assumed that given a sufficient number of data sets the biologically anchored oscillation pattern would predominate.

Our results were comparable between two different university laboratories but as one important finding it is obvious that even though more than 500,000 patients have been included in one data set, sample size was still a limiting factor. The initial exclusion of repeated measurements reduced the size of our data sets by more than 75%. Further, glucose values at night comprised only 10% of all values. Thus, in general more than two million raw data from one site are needed to perform such data mining approaches on hourly glucose values over a 24 h period.

QR analyses were performed, since biological variation may be modulated or differ between different health states or subgroups, such as age classes or gender. The resulting mathematical model is basically compatible with the time-dependent, biologically anchored proportion of the variability of glucose measurement within a population. This type of variation can otherwise only be determined under experimental conditions, i.e., as part of a “constant routine protocol” with serial measurements on healthy volunteers. Acrophases for glucose experimentally determined under such conditions are localized between 23:30 and 00:30 h [18]. According to the median glucose values of our own data (Table 2), the acrophases (t ₀) ranges between 00:16 and 02:14 h. This phase shift would correspond to the time of sample transport.

Notably, the location of the acrophases seemed to be gender-specific in our populations. The circadian rhythm of the glucose level in turn is primarily coupled to the wake–sleep rhythm and does not necessarily follow the course of day and nighttime [3]. Therefore, the phase delay, i.e., later peak, in men might be explained by differences between the sleeping times of men and women, as recently reported in an evaluation of more than 10,000,000 data sets [19]. Strictly interpreted, this literally would imply that before taking a blood sample the time of falling asleep and waking up needs to be recorded in order to have an orientation about the individual cosinor-function.

Further, QRA provided gender and age-specific trajectories at different concentration levels. The shape, i.e., mainly the amplitudes of the respective anticipated time course differed between subgroups. Thus, oscillations of glucose values may not be sufficiently described by a unique so-called “cosinordesm” but comprises several distinct “rhythmodesms”. While no age dependency was observed within each quantile a correlation exists between plasma glucose and the respective amplitudes in a concentration dependent manner in relative and absolute term. This finding is relevant insofar as it indicates that the biological variability within the reference interval may be not strictly linear.

The extend of daily variation can be discussed regarding the overall variation within 24 h or restricted to specific monodesms such as the morning time when blood samples are usually collected. Noteworthy, within the time frame from 06:00 to 10:00 h the intrinsic oscillation of glucose levels reaches its inflection point, thus, the highest change of concentration per time occurs. Considering the increasing demands on analytical accuracy, the observed average change of 4% during this time span could be of relevance. Reported measures of biovariability such as within-subject and between-subject variation tended to be higher [20], [21] which is consistent with the assumption that additional sources of variation exist.

Within-24 h-variation that may be assigned to regular, i.e., physiological, oscillations ranged between ±5.9 and ±9.3% within the interquartile range of the observed glucose values. This amount may therefore be an estimate for the portion of unalterable, i.e., baseline variability to be considered when random glucose values are evaluated.

Referring to the impact of daytime-dependent variations on clinical decision limits provided for random glucose samples we observed a four-fold higher occurrence of cases above 11.1 mmol/L at night. However, we cannot draw any conclusions from this observation because there was no link between our findings and patient records.

Conclusions

The differences between upper RL estimated at day and nighttime exceeded the permissible analytical uncertainty. Because of the observed diurnal variations, indirectly estimated upper RLs of random glucose concentrations should be stratified for sample collection during early morning (e.g., 02:00–06:00 h am) and late morning (e.g., 06:00–12:00 h am). This is particular important if directly determined upper RLs are compared with indirectly estimated upper RLs. Directly determined upper RLs are usually derived from samples collected during late morning. We further propose to extent our approach to other measurands with known or suspected daytime-dependent variations [22].