How this instrument works
The Charlson Comorbidity Index scores nineteen conditions across four weight tiers. Ten conditions carry one point each: prior heart attack, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic lung disease, connective tissue disease, ulcer disease, mild liver disease, and diabetes without organ damage. Six carry two points: hemiplegia, moderate or severe kidney disease, diabetes with organ damage, any non-metastatic tumor, leukemia, and lymphoma. Moderate or severe liver disease carries three points. Metastatic solid tumor and AIDS each carry six. An age term adds a further point for every decade lived past forty: nothing below fifty, one point for the fifties, two for the sixties, three for the seventies, four for the eighties, five at ninety and beyond. Every present condition plus the age term simply add together.
Charlson, Pompei, Ales, and MacKenzie built the original nineteen-condition list in 1987, following patients admitted to a New York hospital and checking which diagnoses independently predicted death within a year; the weights come straight from that statistical fit. The age term arrived seven years later — Charlson, Szatrowski, Peterson, and Gold studied two hundred and twenty-six surgical patients in 1994 and found that each added decade of age carried roughly the same mortality risk as one more comorbidity point, so age was folded into the same total using an identical point scale rather than reported as a separate figure.
The published tool was built to summarize baseline health across groups — matching a trial's treatment and control arms, or comparing surgical outcomes between two hospitals with different patient mixes — not to forecast any single patient's future. A chart review can miss a condition nobody coded, and the index treats a diagnosis as strictly present or absent, with no allowance for how well or poorly it happens to be controlled. Two people sharing an identical total can still carry very different real risk.
- Toggle each one-point condition — from prior heart attack through diabetes without organ damage — to Yes wherever it applies.
- Toggle the two-point fields — hemiplegia, moderate or severe kidney disease, diabetes with organ damage, tumor, leukemia, lymphoma — the same way.
- Set moderate or severe liver disease (three points) if it applies.
- Set metastatic solid tumor or AIDS (six points apiece) if either applies — these two conditions dominate the running total.
- Enter Age in years; the age points calculate automatically and the total reads out below.
Worked example — three 1-point conditions at age 65
A 65-year-old with a history of heart attack, chronic lung disease, and diabetes without organ damage — three conditions at one point each — totals 3 from comorbidities alone. Age 65 sits in the 60-69 decade, which adds 2 age points. The running total closes at 3 + 2 = 5.
Set every comorbidity switch back to No for a 45-year-old with no diagnosed conditions: age points don't start accruing until 50, so this record totals 0, the floor of the index.
Now compare a single metastatic solid tumor, worth 6 points on its own, carried by an 85-year-old. That's 6 for the tumor plus 4 age points (the 80-89 decade), closing at 10 — nearly double the first record, despite that patient logging three separate diagnoses against this one's single entry. That gap is the whole point of the weighting scheme: it isn't built to reward or punish how many boxes get checked, only how much each checked box has historically cost.
Questions
Why do just two conditions outweigh six ordinary ones combined?
Because the 1987 derivation cohort found that a metastatic solid tumor or AIDS predicted death within a year far more strongly than any single 1-point condition — strongly enough that either one alone, at 6 points, equals six 1-point diagnoses stacked together. The weights aren't a subjective severity ranking; they're a direct readout of how much each condition moved the mortality statistics in that original study.
Where did the age adjustment come from, and is it part of the original 1987 index?
No — the 1987 paper by Charlson, Pompei, Ales, and MacKenzie in the Journal of Chronic Diseases defined only the nineteen weighted conditions, with no age term at all. The age points were added seven years later, in a 1994 paper by Charlson, Szatrowski, Peterson, and Gold in the Journal of Clinical Epidemiology, after they found age contributed mortality risk on roughly the same scale as one comorbidity point per decade past forty. Citing 'Charlson Comorbidity Index' without specifying which paper can blur which finding belongs to which study.
What does a score of 0 actually mean?
It means none of the nineteen listed conditions are present and the patient hasn't yet crossed into an age band that adds points — under 50, specifically. It is not a claim of perfect health; conditions outside the nineteen-item list, however serious, simply don't move this particular number.
Can this index predict what happens to one specific patient?
Not reliably. It was built and validated for comparing groups — adjusting for baseline disease burden across a study cohort or a hospital's patient population — where averaging over many people smooths out individual variation. Applied to a single chart, it tells you how much recorded disease that person carries relative to the derivation cohort, not their personal odds.
Does the index distinguish mild disease from severe disease within one category?
No. Each of the nineteen items is a yes-or-no flag; a barely-controlled case of a condition scores identically to a severe one. That flattening is a known tradeoff for keeping the tool fast to complete from a chart review, and it's worth remembering when comparing two patients whose totals happen to match.
Is the Charlson index still used in current research?
Yes, widely — it remains a standard way to statistically adjust for comorbidity burden when comparing outcomes across studies, risk-stratifying before major surgery, or describing a cohort's baseline health in a published paper, decades after its original derivation.
References
- Charlson ME, Pompei P, Ales KL, MacKenzie CR, 1987 — J Chronic Dis (ScienceDirect)
- Charlson M, Szatrowski TP, Peterson J, Gold J, 1994 — J Clin Epidemiol, age adjustment (PubMed)
Read this first: This instrument computes a screening figure from population formulas — it is not a diagnosis, and it cannot see the whole picture a clinician can. Use it to inform a conversation, not to replace one.