Age- and gender-specific trends in glucose dysregulation: insights from everight diagnostics, Owerri, South East Nigeria

Introduction

Type 2 diabetes mellitus and impaired fasting glucose are rising across low-and middleincome countries, with urbanizing populations in Sub-Saharan Africa (SSA) experiencing particularly steep increases in prevalence and associated complications (Atun et al., 2020). The International Diabetes Federation reports elevated population-level burdens and a persistent diagnosis and treatment gap, while WHO surveillance highlights increasing ageLocal analyses and systematic reviews from SSA documented substantial heterogeneity in dysglycemia prevalence that is shaped by urbanization, obesity trends, and inconsistent screening practices (Oguoma et al., 2021; Walker et al., 2023). A critical methodological consideration in these studies is the distribution of fasting blood glucose (FPG). FPG data from both clinical and population samples are frequently right-skewed with heavy tails, meaning a minority of individuals exhibit extreme hyperglycemia (Peer et al., 2023) [12] . Because fasting plasma glucose (FPG) distributions in clinic and population samples are frequently right-skewed with heavy tails, single-measure parametric summaries can underappreciate the burden of extreme hyperglycemia and complicate inference; consequently, careful distributional assessment and appropriate non-parametric or transformed modeling are essential for accurate estimation and policy translation (Worede et al., 2021; Mayega et al., 2022) [15,10] . Consequently, careful distributional assessment and the use of non-parametric or transformed data modeling are essential for accurate estimation and the development of effective public health strategies. This study examines age-and gender-specific trends in FPG among 1,000 diagnostic attendees at Everight Diagnostics, Owerri, South East Nigeria, characterizing distributional properties, demographic associations, and the predictive role of age and sex, with the aim of informing locally relevant screening thresholds and resource allocation for early detection.

Materials and Method Study Design and Population

This cross-sectional study was conducted on 1,000 participants attending Everight Diagnostics, Owerri, South East Nigeria, in 2025. Participants ranged from 1 to 106 years of age, with a balanced gender distribution (51.2% male, 48.8% female). Inclusion criteria required complete records for age, gender, and fasting plasma glucose (FPG), while samples with missing or invalid data were excluded.

The study period was from January 2024 to September 2025 and examined FPG trends in individuals who utilized the diagnostic facility

Sample Collection and Preparation

Venous blood was collected after a minimum of eight hours of fasting into fluoride-oxalate tubes to inhibit glycolysis.

Samples were centrifuged at 1,500-2,000 × g for ten minutes within 30 minutes of collection, and plasma was separated. Samples were analyzed immediately or stored at -20 °C for short-term preservation, minimizing freeze-thaw cycles to prevent glucose degradation.

Fasting Blood Glucose Measurement FPG was measured using the Beckman Coulter DxC 700 AU clinical chemistry analyzer with manufacturervalidated glucose reagents. Calibration was performed at the start of each analytical run using Beckman Coulter calibrators spanning the clinical glucose range, and multilevel quality control was included in every run to ensure precision and accuracy (Bush et al., 2019) [4] . Analytical validation included intra-and inter-run precision, linearity, method comparison with a reference assay, and assessment of interference from hemolysis, icterus, and lipemia (Wang 2020). FPG values were reported in mg/dL and classified according to American Diabetes Association thresholds: normal (<100 mg/dL), impaired fasting glucose (100-125 mg/dL), and diabetes (≥126 mg/dL).

Data Processing and Statistical Analysis

Fasting plasma glucose (FPG) Levels

The mean fasting plasma glucose (FPG) level among participants was 113.25 ± 46.93 mg/dL, with values ranging from 72.4 to 507 mg/dL. The distribution of FPG levels was positively skewed (skewness = 4.13) and leptokurtic (kurtosis = 20.33), suggesting that a few individuals had extremely high glucose readings (Figure 2). The median FPG was 100.85 mg/dL, while the modal FPG was 92.7 mg/dL (Table 3).

Normality Assessment

The distribution of fasting plasma glucose (FPG) was assessed for normality. The Shapiro-Wilk test, indicated a significant deviation from normality, W = 0.76, p < .001.

Similarly, the Kolmogorov-Smirnov test supported this finding, z = 0.19, p < .001. These results suggest that FPG values were not normally distributed in the across the study participants (Table 3).

Gender Differences in FPG Levels

Differences in FPG between males and females were examined (Table 3). Levene's test confirmed homogeneity of variances across gender groups (F = 0.0013, p = 0.9713). Independent samples comparisons showed no significant difference in FPG levels between males (M = 112.92 mg/dL, SD = 50.45) and females (M = 109.50 mg/dL, SD = 50.34), as evidenced by both the Welch t-test (t = 1.1289, p = 0.2592) and the Mann-Whitney U test (U = 125,978, p = 0.3875).

Associations with Categorical Variables

Chi-square tests for associations between FPG categories (normal: <100 mg/dL; impaired fasting glucose: 100-125 mg/dL; diabetes: ≥126 mg/dL) and demographic variables (Table 4). No significant association was observed between FPG category and gender (χ²(2) = 2.2949, p = 0.3176). In contrast, a significant association was found with age group (Table 5) (χ²(6) = 26.6934, p = 0.0002), indicating a smallto-moderate effect Note. χ²(6) = 26.6934, p = 0.0002.

Correlation with Age

Bivariate correlation analyses revealed a weak but statistically significant positive association between age (in years) and FPG levels. Pearson's correlation coefficient was r = 0.1341 (p < 0.001), and Spearman's rank correlation was ρ = 0.1307 (p < 0.001), suggesting that FPG tends to increase modestly with advancing age.

Age Group Comparisons

One-way ANOVA revealed significant differences in mean FPG across age groups (F(3, 496) = 8.8068, p < 0.0001).

The non-parametric Kruskal-Wallis test confirmed this finding (H(3) = 21.9935, p = 0.0001).

Post-hoc pairwise comparisons using Tukey's Honestly Significant Difference (HSD) test are presented in Table 6. The 60+ age group exhibited significantly higher FPG levels than the 0-19, 20-39, and 40-59 groups. The 40-59 group had higher FPG than the 20-39 group, though this difference did not reach statistical significance at the 0.05 level (p = 0.099).

Regression Analysis

A multiple linear regression was conducted to examine whether age and gender predicted fasting plasma glucose (FPG) levels. The dependent variable was FPG, while the independent variables were age and gender (coded numerically).

Model Summary

The overall regression model was statistically significant, F (2, 997) = 6.70, p = .001, but explained only a small proportion of the variance in FPG, R² = .013, adjusted R² = .011, indicating limited explanatory power. Age was a significant predictor of FPG, with higher age associated with higher FPG (B = 0.34, p < .001). Gender did not significantly predict FPG (B = 0.98, p = .741). Overall, the model accounted for only 1.3% of the variance in FPG, suggesting that other factors not included in this model may contribute more substantially to fasting plasma glucose levels. Note: B = unstandardized regression coefficient; SE B = standard error of B; CI = confidence interval.

Logistic Regression Analysis

To examine predictors of diabetes status (binary: 0 = non-diabetic [<126 mg/dL], 1 = diabetic [≥126 mg/dL]), a binary logistic regression model was fitted with age (in years) and gender (0 = female, 1 = male) as independent variables (Table 8). The model was statistically significant (likelihood ratio χ²(2) = 9.50, p = 0.009), indicating that the predictors improved fit over the null model. However, the pseudo R² (McFadden) was low (0.012), suggesting that other factors contribute more substantially to diabetes risk. Age significantly predicted diabetes status (β = 0.0177, p = 0.002), with an odds ratio of 1.018 (95% CI [1.006, 1.030]). This indicates that each additional year of age is associated with a 1.8% increase in the odds of being diabetic, controlling for gender. Gender was not a significant predictor (β = -0.0081, p = confirming no association between gender and diabetes risk in this sample. The model converged successfully, and standard errors were based on maximum likelihood estimation.

Multinomial Logistic Regression Analysis

A multinomial logistic regression was performed to predict FPG category (Reference: Normal). The full model significantly improved fit over the null model, likelihood ratio χ²(4) = 42.42, p < .001, with a McFadden R² of 0.021.

Age significantly increased the odds of being in the IFG category (OR = 1.022, p < 0.001) and the Diabetes category (OR = 1.031, p < 0.001). Males had 45.6% higher odds of being in the IFG category than females (OR = 1.456, p = 0.007), but no significant gender difference was found for the Diabetes category (OR = 1.146, p = 0.488) (Table 9).

Discussion

The study population consisted of [1] . The demographic pattern supports calls for age-stratified screening policies advocated by international disease burden reports (IDF, 2021; WHO, 2022). The distribution of FPG was markedly non-normal, characterized by significant positive skewness and kurtosis. This indicates that while most of the population had glucose levels clustered in the normal to pre-diabetic range, a substantive minority exhibited extreme hyperglycemia, dragging the mean upward. This pattern is a hallmark of populations undergoing epidemiological transition, where a high burden of undiagnosed or poorly managed dysglycemia exists alongside a generally normoglycemic population (Peer et al., 2023; Mayega et al., 2022) [12,10] . The profound deviation from normality, confirmed by formal statistical tests, validates the use of non-parametric analytical techniques in this and similar SSA datasets to avoid biased estimations of effect sizes and prevalence (Worede et al., 2021) [15] . The sampled cohort's fasting plasma glucose (FPG) had a mean of 113.25 ± 46.93 mg/dL, a median of 100.85 mg/dL, conspicuous maximum extremes (as high as 507 mg/dL). The combination of the mean > median, strong right tail, and heavy tails, indicate there is a minority of individuals with significant hyperglycemia that influenced the measure of central tendency and signals a high burden of undiagnosed and/or poorly managed dysglycemia amongst the individuals sampled. These types of distributional evidence mirror urban-sub-Saharan Africa diagnostic cohort and meta-analyses that show elevated mean fasting plasma glucose (FPG) and wide dispersion due to late presenters and a variety of risk exposures. This is consistent with published syntheses, regionally focused, which show a broad prevalence of prediabetes/diabetes in urban-sub-Saharan Africa (Oguoma et al., 2021; Walker et al., 2023) [11, 13] .

Results of Formal Normality Tests (Shapiro-Wilk W = 0.7629, p < .001; Kolmogorov-Smirnov z = 0.1877, p < 001) revealed a substantial deviation from Gaussian assumptions for FPG. The statistical implication is that applying parametric tests, which rely on normal residuals or symmetric distributions, are likely biased; therefore, nonparametric alternatives (i.e., Mann-Whitney, Kruskal-Wallis) and use of robust variance estimators is warranted and enhances the inferential validity. Empirically, it has been established that many SSA glucose datasets exhibit similar deviations from normality regarding the empirical distributions of glucose measurements, and studies have provided evidence that inappropriately applying parametric models to skewed glucose data can lead to underestimating the prevalence of prediabetes or misestimation of effect sizes (Worede et al., 2021) [15] .

A key finding was the absence of a significant difference in mean FPG between males and females. This finding reflects there were no substantial sex differences in glycemic burden among urban diagnostic attendees, and as such genderneutral screening and prevention policies could be warranted for this context. This null finding adds support to urban SSA analytic work, whereby sex differences in mean glycemia have been shown to be minor, but past studies, in particular, studies examining older, or rural females, have described modestly elevated female levels in relation to varying care-seeking or menopausal transition (Walker et al., 2023) [13] This aligns with several recent urban SSA studies that report minimal sex disparities in overall glycemic levels, suggesting that in urban environments, traditional genderspecific risk factors may be converging due to shared lifestyle and environmental exposures (Adeleye et al., 2020) [2] . Thus, this finding is consistent with the literature showing contextual heterogeneity: sex differences are small or nonexistent among some urban samples while differences can be present in other populations based on the historical and current age structure, and level of urban/rural status. However, the multinomial logistic regression revealed a more nuanced picture: males had 46% higher odds of having IFG than females, even after controlling for age. This suggests that the early dysglycemic phase might be a critical window where gender differences are more pronounced, possibly reflecting higher-risk behaviors (such as dietary patterns, alcohol use, or physical inactivity) among men before the progression to frank diabetes, where physiological factors may dominate and diminish the gender gap (Kibirige et al., 2021) [9] . This finding underscores the potential for targeted pre-diabetes screening and lifestyle intervention specifically for adult males.

The most robust and consistent finding across all analyses was the strong, graded association between advancing age and worsening glucose control. The prevalence of diabetes was over 17 times higher in the 60+ age group compared to the 0-19 age group. Correlation and regression analyses consistently showed that age, though a statistically powerful predictor, explains only a small portion of the total variance in FPG. This is a common observation in epidemiological studies and highlights that age acts as a proxy for cumulative beta-cell function decline and lifelong exposure to other risk factors (Chatterjee et al., 2023) [5] . [15,8] . Multinomial and logistic models indicated that age raised odds of impaired fasting glucose and diabetes; a small male excess was found in impaired fasting glucose (OR ≈ 1.456) but not established diabetes. Such data suggest differences between sexes are more evident with early dysglycemic stages and might even be subtle transitions (potentially indicative of the behavioral-or occupational-exposure) before euglycemia transitions into frank diabetes where sexes are similar in this sampled population. The models' low pseudo-R² values (McFadden ≈ 0.01-0.02) suggests there are still missing predictors; particularly measures of adiposity and lifestyle that, once included, generally improve model performance (Kang et al., 2021; Worede et al., 2023) [15,8] . The significant age gradient, particularly the marked increase in dysglycemia from age 40 onwards, has direct implications for public health policy. Our data strongly support the implementation of systematic, population-based screening for glucose dysregulation starting at age 40, as recommended by various guidelines for high-risk populations (IDF, 2021). The combination of age-targeted screening, coupled with specific outreach for men in the prediabetes stage, could form a cost-effective strategy for early detection and intervention. Community-based programs focusing on weight management and physical activity, integrated with the measurement of other cardiometabolic risk factors like hypertension and dyslipidemia, are essential to curb the rising tide of T2DM in this region (Gouda et al., 2023).

References

1. Adeboye AA, Oladapo OO, Adebayo OM. Profile of patients accessing private laboratory services in a Nigerian urban setting: Implications for public health planning. J Community Health. 2022;47(3):512-520.
2. Adeleye OO, Owolabi MO, Ogunniyi A. Gender and urban-rural differences in the prevalence of diabetes and impaired fasting glucose in Nigeria: The Ibadan Menopause Study. Diabetes Res Clin Pract. 2020;166:108-115.
3. Atun R, Davies JI, Gale EAM, Bärnighausen T, Beran D, Kengne AP, et al. Diabetes in sub-Saharan Africa: from clinical care to health policy. Lancet Diabetes Endocrinol. 2020;8(5):418-448.
4. Bush VJ, Smola C, Schmitt P. Evaluation of the Beckman Coulter DxC 700 AU chemistry analyzer. Pract Lab Med. 2019;18:148-156.
5. Chatterjee S, Khunti K, Davies MJ. Type 2 diabetes. Lancet. 2023;401(10391):1803-1820.
6. Gouda HN, Charlson F, Williams S, Nisbett N. The double burden of malnutrition and associated cardiometabolic risk factors in sub-Saharan Africa: A systematic review and meta-analysis. Lancet Glob Health. 2023;11(2):e235-e245.
7. International Diabetes Federation. IDF Diabetes Atlas. 10th ed. 2021.
8. Kang SY, Kim YS. Relationships between fasting glucose levels, lifestyle factors, and metabolic parameters in Korean adults without diagnosis of diabetes mellitus. J Diabetes. 2021;14(1):52-63.
9. Kibirige D, Akabwai GP, Sseguya W. Prevalence and correlates of prediabetes and diabetes mellitus among adults in Uganda: A national population-based survey. BMJ Open Diabetes Res Care. 2021;9(1):e002133.
10. Mayega RW, Etajak S, Rutebemberwa E, Tomson G, Kiguli J. A skewed distribution of cardiovascular risk factors at a primary care clinic in rural Uganda: A cross-sectional study. Glob Health Action. 2022;15(1):1-12.
11. Oguoma VM, Nwose EU, Richards RS. Prediabetes in sub-Saharan Africa: A systematic review and metaanalysis. Front Public Health. 2021;9:762-778.
12. Peer N, Balakrishna Y, George J. The distribution of fasting plasma glucose and the implications for diabetes screening in South Africa: A systematic review and meta-analysis. S Afr Med J. 2023;113(4):25-32.
13. Walker RJ, Thorgerson AM, Yan A, Williams JS, Campbell JA, Dawson AZ, et al. Prevalence and correlates of pre-diabetes in Sub-Saharan Africa: Country comparisons and implications. BMC Public Health. 2023;23(1):222-236.
14. Wang Y. Hemolysis interference assessment on the DxC 700 AU chemistry analyzer. Clin Biochem. 2022;103:1-6.
15. Worede A, Alemu S, Gelaw YA, Abebe M. Prevalence of prediabetes and associated factors among adults in Debre Tabor Town, northwest Ethiopia, 2019: Community-based cross-sectional study. BMC Res Notes. 2021;14:280-291.
16. World Health Organization. Global report on diabetes. Geneva, Switzerland: WHO; 2022. Available from: https://www.who.int/health-topics/diabetes
17. Zhou B, Lu Y, Hajifathalian K, et al. Worldwide trends in diabetes since 1980: A pooled analysis of 751 population-based studies with 4.4 million participants. Lancet. 2023;401(10392):1905-1916.