Number needed to treat to harm (NNTH) analysis of impact of underlying cardiovascular factors on risk of abacavir-related heart attack

Simon Collins, HIV i-Base

JD Kowalska from the D:A:D study presented a model looking at the ‘number needed to treat to harm’ (NNTH) in order to help interpret the clinical importance of the 90% increased relative risk (RR=1.90), previously reported between abacavir use and the risk of heart attack. [1, 2]

This estimate changed depending on an individuals underling cardiovascular risk. The underlying risk was calculated based on 5-year Framingham score [3] and the relative rate was assumed to remain constant across the range of underlying risk of MI.

NNTH was calculated as:

1 ÷ (underlying risk of MI x 1.90 – underlying risk of MI)

The lowest NNTH values were observed in the high risk group, while the most dynamic changes in NNTH is in the low risk group, showing an exponential relationship between NNTH and underlying risk of MI. The NNTH dropped steeply from 185 with an underlying risk of MI of 0.6% to only 5 when underlying MI risk is 20%.

The importance of additional risk factors was illustrated by starting with a low risk patient (0.1% risk of MI); in this case, 1111 patients would need to be treated before seeing an abacavir-related MI. When two unfavourable risk components are present the NNTH drops to around 100 for most pairs, except smoking and low HDL, for which NNTH drops to 69. When all risk factors are unfavourable, in patients with 15% underlying CVD risk, only 7 people would need to be treated with abacavir to see a treatment associated MI. (See Table 1).

Table 1: Change in factors contributing to underlying risk

Underlying risk of MI in 5 years 5-year risk of MI (%) NNTH
Low risk profile (40 year old man with none of the risk factors listed below) 0.1 1111
Impact of addtional single risks
If total cholesterol 240 mg/dL (6.2 mmol/L) 0.2 555
If diabetes 0.2 555
If ECG-LVH 0.2 555
If sBP 160 mmHg 0.3 370
If HDL 35 mg/dL (0.9 mmol/L) 0.3 370
If smoking 0.4 277
Impact of multiple risks
If HDL and total cholesterol unfavourable 0.8 138
If smoking and diabetes 1.1 101
If smoking and total cholesterol unfavourable 1.0 111
If smoking and sBP 160 mmHg 1.3 85
If smoking and HDL unfavourable 1.6 69
If smoking and lipids unfavourable 3.1 35
If all unfavourable combined (excluding ECG-LVH) 10.1 11
If all unfavourable combined (including ECG-LVH) 15.0 7


This model supports the conclusions from both D:A:D and SMART studies to caution against using abacavir in patients with high underlying cardiovascular risk.

It also demonstrates the potential clinical impact from reducing other risk factors where alternative treatment options are not available.


  1. Kowalska JD et al. Relation between adverse effects of ARV treatment and underlying risk in number needed to treat to harm (NNTH) – myocardial infarction and abacavir use . 9th International Congress on Drug Therapy in HIV Infection. 9-13 November 2008, Glasgow. Abstract O313.
  2. Sabin C, et al. Use of nucleoside reverse transcriptase inhibitors and risk of myocardial infarction in HIV-infected patients enrolled in the D:A:D study: a multi-cohort collaboration. Lancet 2008; 371 (9622): 1417-26.
  3. Anderson KM, et al. Cardiovascular disease risk profiles. Am Heart J 1991;121: 293-298.

Links to other websites are current at date of posting but not maintained.