Results
Search Results
The systematic literature search yielded 5770 articles. Figure 1 shows the flow diagram for the identification process. Briefly, 5498 articles were excluded based on title and abstract. In the next step, 217 articles were excluded based on full-text reviewing and language criteria. The remaining 55 studies were identified in full-text; 41 studies were excluded based on non-prospective study design (n = 30 cross-sectional, n = 11 case-control). The hand search did not result in any additional articles. The email alerts yielded in 2 additional studies. Sixteen studies were eligible for quality assessment and ranked according to their sum scores (see online Appendix Supplement 3). One study was graded as "poor" and therefore excluded. Finally, 15 eligible studies were included, of which 14 studies were grated as "good" and 1 as "moderate" quality.
(Enlarge Image)
Figure 1.
Flow diagram of the identification process for eligible studies. *Languages: Russian (n = 11), Italian (n = 6), Japanese (n = 6), Polish (n = 6), Spanish (n = 4), French (n = 3), Bosnic (n = 1), Finnish (n = 1), German (n = 1), Romania (n = 1), Servic (n = 1), Czech (n = 1). † Animal studies (n = 2), in vitro studies (n = 1). ‡ Patients with dialysis (n = 25), chronic kidney disease (n = 12), primary hyperparathyroidism (n = 4), parathyroidectomy (n = 4), other (n = 2).
Study Characteristics
Of the 15 articles, 12 investigated the relation between PTH and total CVD events of which 8 investigated fatal CVD events. Table I represents the main characteristics and results of the included studies on CVD events. Multiple articles were based on the same cohort—Cardiovascular Health Study and Uppsala Longitudinal Study of Adult Men; we therefore included only the more comprehensive studies to avoid overlap in participants. This resulted in 10 studies for the meta-analysis for PTH and total CVD events. Seven studies investigated PTH and fatal CVD events and 3 studies investigated PTH and non-fatal CVD events of which 2 studies reported both total CVD events and non-fatal CVD events. The total number of events was 2745 for total CVD of which 1466 were fatal. The studies that investigated specifically non-fatal CVD events compromised 1018 events. Four prospective studies focused on intermediate outcomes; delta blood pressure, incident hypertension, and left ventricular mass (LV mass), a measure of cardiac remodeling.
In all studies, PTH measurements were determined to study relationships with chronic diseases in older participants and participants were not selected based on PTH or 25-hydroxyvitamin D concentrations. The population sizes of the studies, all published between 2005 and 2012, varied between 148 and 5569. The mean study duration ranged between 2 and 14 years. Overall, 12 different cohorts were used of which 3 included only men. Six cohorts were conducted in Europe and 6 in the United States, all primarily in whites (0%-14% black), except 1 multi-ethnic study; mean age ranged between 55 and 75 years, except for 1 study, 30 to 70 years. CVD events were defined based on International Classification of Diseases (ICD)-9 and ICD-10 codes in 6 of 12 cohorts. All studies adjusted for estimated glomerular filtration rate or creatinine—measures of kidney function—in multivariate models, except 1. Three studies excluded participants with parathyroid disorders at baseline. Furthermore, 3 studies performed sensitivity analyses for participants with high PTH and high calcium concentrations — indicative for primary hyperparathyroidism — which did not change the results (prevalence between 1% and 11%). Other studies did report the possibility that some of the participants could have primary hyperparathyroidism, or did not report any information about parathyroid disorders. Risk estimates (HRs/ORs), regression coefficients, and 95% CI were reported for PTH quartiles (Q4 vs Q1) or dichotomous cut-off values of PTH. The funnel plot (Figure 2) for studies of PTH and total CVD events shows reasonable symmetry at the top of the funnel plot and a little asymmetry at the bottom, which suggest some evidence of publication bias for smaller studies.
(Enlarge Image)
Figure 2.
Funnel plot for PTH and total CVD events. Each square indicates 1 study with its standard error indicating the weight of the study and its relative risk. The dotted lines represent 95% CI to visualize the symmetry around the pooled estimate; symmetry indicates no evidence of publication bias.