The rise and fall of bilirubin and alkaline phosphatase

The rise and fall of bilirubin and alkaline phosphatase

Why, after an episode of prolonged biliary obstruction, or after severe forms of acute liver injury, does the alkaline phosphatase (ALP) rise in a delayed fashion? And why do the ALP and total bilirubin often take a while to normalize?

We’ve all heard (or said) this before. Today, let’s get to the bottom of why this happens!

First, let’s review bilirubin metabolism.

We can keep this brief and focus on the relevant aspects. For a more in-depth discussion of bilirubin metabolism, refer to this excellent review by Johan Fevery.

Also check out my previous infographic on the basics of the liver chemistries!


  1. Heme breakdown (80% of which comes from senescent erythrocytes) creates biliverdin
  2. Biliverdin → converted to un-conjugated bilirubin (UCB)
  3. UCB is poorly water soluble = requires albumin to travel through plasma
  4. Albumin-bound UCB → liver → enters hepatocytes → dissociates from albumin
  5. In hepatocyte → bilirubin is conjugated by uridine 5’-diphospho-glucuronosyltransferase 1A1 (UGT1A1) [notably, this is the enzyme deficient in Gilbert’s and Crigler-Najar]
  6. Conjugated bilirubin (CB) = excreted in an ATP-dependent fashion across the hepatocyte canalicular membrane into bile via the multidrug resistance-associated protein 2 (MRP2)
  7. However, a proportion of CB is actually secreted back into the sinusoidal blood via the MRP3 transporter, where it has one of two fates:
    • Most is just taken back up by the hepatocyte via OATP1B1 and OATP1B3 transporters (seems redundant? I agree.)
    • A small percentage remains in plasma and is excreted in urine (which is possible since CB is now water-soluble)

Ok, so in total, bilirubin can exist in 4 forms:

    1. Unconjugated bilirubin bound to albumin
    2. Free bilirubin (not bound to albumin, toxic, causes kernicterus)
    3. Conjugated bilirubin
    4. Delta bilirubin 

Wait…what’s delta bilirubin?

Simply put, delta bilirubin is conjugated bilirubin that becomes irreversibly bound to albumin in plasma.

We initially said that UCB, which is poorly water-soluble, requires albumin to circulate in plasma, right?

We also said that once in the hepatocyte, some proportion of CB is actually secreted back into plasma rather than directly into bile (via MRP3), and then is reuptaken by the hepatocyte (via OATP1B1/B3).

Well, in cases of obstructive jaundice or other episodes of acute liver injury, the ability to excrete CB into bile is impaired. If CB can’t go that way, more of it goes back into plasma and accumulates there. CB that sits in plasma for a long time gets bound to albumin in an irreversible fashion (see here and here and here). This is delta-bilirubin.

There are few characteristics of delta-bilirubin that are important to stress:

  • Binding to albumin is irreversible
  • This means that it stays in plasma (it cannot be excreted in urine/bile while bound to albumin, as it is no-longer water-soluble). 
  • Half-life is similar to that of albumin (17 to 21 days)
  • It is likely non-toxic

Remember, delta-bilirubin is conjugated, so it will be counted in the direct bilirubin measurement in typical laboratories.

Now that you mention it, why do we use the terms “direct” and “indirect” when measuring bilirubin?

This is a bit of a digression, but it is good to know how bilirubin is measured.

There are multiple ways to measure bilirubin, but the terms “direct” and “indirect” come from the van der Bergh reaction (“diazo” reaction).

In the van der Bergh method, the following occurs:

  • A diazo compound (i.e., a compound with two nitrogen atoms) is used to react with the bilirubin present in the blood sample
  • The diazo compound starts to react with both types of bilirubin (UCB and CB) immediately
  • However, it reacts much more quickly (“directly”), aka within minutes, with CB (this is because CB does not have internal hydrogen bonds)
  • The diazo compound takes much longer to react with molecules that do have internal hydrogen bonds, such as UCB (this reaction typically requires an accelerator, such as alcohol) 
  • In sum, in the van der Bergh reaction:
    • The first measurement obtained is the “direct”-acting bilirubin, which is approximated as the amount of CB
    • The second measurement, obtained after the entire reaction is complete, is the total bilirubin
    • Finally, indirect bilirubin, which is approximated as the fraction of UCB, is calculated as the difference between the total and direct bilirubin values

Two important things to know:

    1. “Direct” bilirubin can overestimate the amount of CB

      This is because, as I said initially, the diazo compound actually does begin reacting with both CB and UCB from the start. It just reacts much more slowly with UCB. So while most of the bilirubin being measured in the early first few minutes is CB, some will be UCB as well.

      Usually, the error is small, but as total bilirubin levels rise, the error also increases

    2. Direct bilirubin also measures delta-bilirubin (as it is conjugated and thus reacts in that early part of the reaction)

Okay, enough about bilirubin. What about alkaline phosphatase?

Alkaline phosphatase (ALP) is an enzyme (of the zinc metalloproteinase family).

Importantly, it is produced within the hepatocyte. It resides on the hepatocyte canalicular membrane (the membrane facing the bile ducts).

It is well known that the liver fraction of ALP rises in the setting of cholestasis (e.g., bile duct injury, obstruction, etc). 

Importantly, experimental studies of the physiology of this rise in ALP have shown:

From the above, it should start to make sense why the increase in ALP usually does not occur immediately. The increase in plasma levels will only occur after ALP synthesis has increased AND after the canalicular membrane becomes saturated. 

Also important: the half-life of ALP is approximately 7 days. Therefore, depending on how high it peaks, we can expect its improvement to take up to weeks after the resolution of injury. 

I think I get it. Can you summarize for me?

To put it all together:

  1. Delta-bilirubin is a form of CB that is irreversibly bound to albumin, and it forms when CB sits in plasma for an extended period of time (i.e. in ongoing biliary obstruction or liver injury that causes an inability to excrete conjugated bilirubin)
  2. Thus, delta-bilirubin’s half-life is similar to albumin, and it can take weeks to months to completely disappear
  3. ALP rises because hepatocyte bile acid accumulation (not necessarily related to bilirubin metabolism) causes increased synthesis of the enzyme 
  4. ALP has a half-life of 7 days and thus can also take weeks to improve after a causative injury
Written by:

Transplant Hepatologist, University of North Carolina

Related content