PTH is an independent risk factor for mortality in patients with CKD on maintenance dialysis [12, 9, 5, 2, 8, 6, 4, 1, 7, 3, 10, 11]. The current study extends observation in these earlier studies providing new data from a middle-income country and highlighting the risk of mortality during the first 90 days of therapy. Patients who started dialysis with PTH < 150 pg/mL had an 89% higher risk of mortality in an adjusted model. Age, serum albumin and the first dialysis in a Hospital setting also were merit as risk factors for mortality. Taken together, our results identified that patients’ profiles, specifically public health dependence, aging and malnutrition might have contributed to this result.
Patients included in this analysis are relatively young, and, although we have no information on shared decision-making to start dialysis, it is unlikely that low life expectancy due to age and frailty are concerns. Even in this scenario, the observed mortality rate was 29.7% during the study period. A previous systematic review that analyzed the prediction of risk of death for patients starting dialysis included 36 studies in patients on incident patients on dialysis described a mortality rate between 6.1% and 55.5% [17]. All studies in this review included the first 90-day mortality.
As the first 90 days after starting dialysis carry a higher risk of mortality, information on this phase is of utmost importance. Mortality within 90 days has been described as 8.6% [18], 10.5% [19] and 12.3% [20] in older patients. We found a lower mortality rate in this period, 6.7%, but in a younger population. Even large epidemiological registries such as DOPPS are limited to describing patients with ≥ 3 months follow-up. Indeed, the UK Renal Registry called attention that quality assurance is based on populations depleted of those who died in the first 90 days of therapy [21]. Whether the recognition of predictive factors of mortality in the first 90 days of therapy would allow therapeutical changes to reverse the prognosis is unknown.
Previous studies that focused on evaluating fibroblast growth factor 23 in patients starting dialysis, showed no relationship between PTH and mortality [22, 23]. To the best of our knowledge, there is only one previous study conducted to evaluate the mortality rate in 424 incident patients on hemodialysis in our country [8]. PTH levels of patients who died in 1 year were 146 pg/mL against 165 pg/mL in those who survived, a non-significant difference. Since PTH was not categorized and there was no specific analysis within the first 90 days of therapy in the above-mentioned studies, comparison with our results was not doable.
Other variables that were associated with early and long-term mortality were age, serum albumin, and first dialysis at the hospital, which are known factors related to mortality in patients on dialysis [24, 25]. We could speculate that low PTH would reflect a malnutrition/inflammation condition. However, PTH < 150 pg/mL remained independently associated with early mortality even after multiple adjustments. Beyond this period, PTH was not associated with mortality in a fully adjusted method. Age, on the other hand, was confirmed as a true risk factor for mortality.
Data from the Dialysis Outcomes and Practice Patterns Study that evaluated PTH in patients initiating dialysis [26] identified that 23.9% and 16.4% of patients had a PTH < 150 pg/mL or PTH > 600 pg/mL, respectively. In contrast, we found 32.0% and 16.6%, respectively, a higher prevalence of individuals with low PTH. There is no plausible explanation for this discrepancy since our patients were younger than in the DOPPS data and had a similar prevalence of diabetes [26]. Unfortunately, there was no survival analysis to compare with our results. We found that patients with PTH < 150pg/mL had a higher mortality risk in the first 90 days of therapy, a result not sustained in 1 and 5 years. Taken together our results show that the mortality in the first 90 days of therapy seems to reflect a nutritional aspect as a risk factor for mortality that is independent of long-term therapy.
In Brazil, all individuals have the right to the Public Health System, which is a national health system called SUS that covers within the country any treatment or medication even in the most complex cases. However, private health insurance is allowed for those who can afford it. We found the mortality rate in 5 years was higher among patients covered by the SUS than others with private insurance despite similar dialytic treatment. This result might reflect the easier access by the patients with private insurance to medical exams, treatment and follow-up by other specialties, extra dialytic medicines, as well as hospitalization when needed (27).
Our study has some limitations. First, the observational design. Second, we have no data on mineral and bone-associated medication. The strengths of this study are the inclusion of incident patients, the analysis of mortality within 90 days after starting dialysis, and the sensitive analysis of changes in the PTH category during the first year of therapy.
Based on our results we can conclude that patients who started dialysis with PTH lower than 150 pg/mL have a higher risk of death during the 90 following days. Neither the PTH level at the beginning of renal replacement therapy nor the change in PTH category after 1 year seems to be merit as a risk factor for 1-year and long-term mortality after adjustments for confounders.