In the early years of my career, we medical relics drew blood, started IVs, sometimes looked at wounds, and definitely used our wooden tongue depressors without first donning nonsterile gloves. We were taught as students to wash our hands before and after examining each patient, though ample surveys showed a chasm of adherence among doctors, who moved from one patient to the next with their hands unwashed.
We protected ourselves when we knew about an active infection, such as hepatitis, and assumed — mostly correctly — that we would not contract an occupational infection otherwise. That world changed in the 1980s when AIDS became the infection that health workers could acquire unwittingly and hepatitis expanded its silent prevalence as testing became more proficient. For well-validated safety reasons, universal precautions rapidly became the expectation, and all people are now assumed to be potential sources of public risk. Most are not, but we have no expedient means of screening, so now we all wear gloves and use gel disinfectant.
Our approach to some common diseases has taken a parallel path. We developed algorithms telling us that certain cancer screenings and immunizations apply to everyone on a set schedule. Automated reminders or electronic health record nudges assure that we regularly check A1c levels in patients with diabetes. We try to juggle universality with specific circumstances, and for the most part, care has become more organized, with fewer omissions of the essentials.
Yet, the bulk of the diseases we confront and the people who have them baffle us with their heterogeneity. Many of our professional societies have suggested that the move of the medical pendulum toward universality has come at the expense of helpful stratification of the people we see, along with our ability to set priorities for addressing the differences among patients with the same condition.
Two recent studies — both population reviews of data collected previously — suggest ways to better sort the risk among large groups with prediabetes and established diabetes. Despite the high prevalence of prediabetes (as much as 50% in some senior populations, depending on the definition you use), a good deal of uncertainty remains about its natural history and progression to diabetes.
Risk in Older Age
Rooney and colleagues captured a cohort of people who had entered a large database as young adults and are now senior citizens. Data from their assessments in 2011 identified a group who met the American criteria for prediabetes, with either an A1c of 5.7%-6.4% or a fasting blood glucose level of 100-125 mg/dL. Was the progression to type 2 diabetes inevitable, likely, or uncommon? Did the means of identifying prediabetes select out the people who needed the most attention?
Baseline abnormalities were quite common, with about 73% of the enrolled participants having one of the two prediabetes criteria. When European criteria for prediabetes were used (A1c of 6.0% and fasting blood glucose of 110 mg/dL), the number of people meeting the criteria plummeted, suggesting that many were at the lower edges of the American definitions.
Of the people identified as prediabetic in 2011, 9% progressed to diabetes either by lab data or by documented use of hypoglycemic agents 6 years later. A slightly higher 13% were euglycemic, and 19% had died during that interval.
All participants were older, but those older than 75 years had the same progression rates to diabetes as those under 75. Gender did not predict progression. However, there was a small but statistically significant progression among Black people. A1c criteria had better predictive value than fasting blood glucose levels, but having both the A1c and the fasting blood glucose in the prediabetes range seemed to enhance the prediction of progression.
Most pertinent among these older patients, whose mortality rates exceeded any changes in glycemia, is that the presence of prediabetes in 2011 had no correlation with mortality during the study interval. Borderline glucose levels or A1c among older individuals did not add much to the crystal ball that most of us would like to have as we assess patients and try to protect their futures.
Racial and Ethnic Differences
In another study, Bancks and colleagues assessed the ethnic composition of patients within five commonly recognized diabetic clinical phenotypes. Many of us see a highly diverse mix of patients. Sometimes that diversity reflects ethnicity and sometimes it reflects a spectrum of illness. The study attempted to define clinical phenotypes, which often have different ethnic distributions.
As part of their effort to stratify risk, the investigators also looked at the risks for coronary disease and chronic kidney disease across different clinical presentations and by ethnicity within each clinical division.
The investigators extracted data from two populations from ongoing studies, one involving American South Asians and the other a more heterogeneous group of Americans with ethnic distributions far different from the general population. They identified five groups of patients who would cause us to focus on a specific variant of diabetes: older age of onset (43%); severe hyperglycemia (26%); severe obesity (20%); insulin users (9%); and onset before age 30 (1%).
For the most part, fewer Chinese Americans and South Asians had obesity. Chinese ancestry predicted an older age of onset. South Asians were disproportionately represented in the group with severe hyperglycemia and had the most impairment of beta-cell function. The risk for kidney involvement or coronary disease correlated more with clinical phenotype than with ethnicity within a phenotype.
Both studies reinforce that diabetes, as with much of medicine, involves components that are universal to everyone but particular to some. From the prediabetes data, we can reassure our older patients that their borderline glucose levels are not yet disease and are more likely to improve spontaneously than to progress to established diabetes. For those with established diabetes, the ethnic study offers less ability to be reassuring, perhaps even risking inaccurate stereotypes.
The initiative of defining heterogeneity has advanced somewhat, but the corollary goal of using those classifications to ease the lives of those at less risk remains elusive. We can identify phenotypes of hyperglycemia as we always have but not subdivide very well who fits into each presentation. We are left with a form of universal management to minimize end organ damage, much as universal precautions for infection control are best applied to everyone, accepting that care will be excessive for some to ensure that it is adequate for all.
Richard M. Plotzker, MD, is a retired endocrinologist with 40 years of experience treating patients in both private-practice and hospital settings. He has been a Medscape contributor since 2012.
Medscape Diabetes © 2021 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Universal Care Algorithms for Diabetes May Not Work for Everyone - Medscape - Aug 27, 2021.