Download Number Needed to Treat: an Important Measure for the Correct

Editorial
DOI: 10.5455/bcp.20150318073223
Number Needed to Treat: an Important Measure
for the Correct Assessment of Clinical Significance
Mesut Cetin1, Selim Kilic2
Bulletin of Clinical Psychopharmacology 2015;25(1):1-3
Number Needed to Treat (NNT) is a concept
first found in papers published by Laupacis et al.
in the New England Journal of Medicine in 1988.
NNT is the average number of patients who need
to be treated to prevent one additional bad
outcome (e.g. the number of patients that need to
be treated for one to benefit compared with a
control in a clinical trial)1. This measure assessing
the clinical significance of any kind of intervention
has since been applied with increasing frequency.
This concept is frequently used in psychiatry,
especially when comparing the efficacy and
effectiveness of drugs and is seen as an important
indicator to demonstrate the advantage of a new
drug over a standard or reference drug and/or
placebo treatment2-5.
In addition to NNT, there are two more
concepts in use for the comparison of drug
efficacy and/or effectiveness namely, “Absolute
Risk Reduction” (ARR) which is the efficacy of
drug 1 – the efficacy of drug 2 and “Relative Risk
Reduction” (RRR) which is the efficacy of drug 1 /
the efficacy of drug 26-8. Thus, ARR measures the
difference in efficacy between two drugs and RRR
the ratio between the efficacies of two treatments.
Given that, of these two measures, RRR may make
the effect size appear higher than it is in reality, it
has been established that using NNT, which is
related to ARR (NNT=1/ARR), allows a more
accurate assessment of the effect size).
In very large samples, the p value, which is
affected not only by effect size but also by sample
size, can be very close to zero (p<0.001) even
though there is no clinically significant difference.
After all, the p value ascertains that the difference
found in the comparison is statistically significant,
i.e., not accidental. It is thus possible that a
statistically significant p value may be wrongly
interpreted as implying that there is a clinically
significant difference. By contrast, the NNT value,
which is unaffected by the sample size affecting
the p value, is a suitable measure to demonstrate
and interpret clinical significance10-11. The closer
the NNT value is to 1, the more effective is the
drug.
On the other hand, drugs can lead to
undesirable side effects in clinical application. To
give some examples of commonly found side
effects, antidepressants can lead to weight gain,
sedation, and sexual side effects2 and
antipsychotics used in patients suffering from
schizophrenia or bipolar disorder can provoke
extrapyramidal side effects, weight gain, an
increase in prolactin or a QTc prolongation in the
ECG12-14. To quantify the undesirable side effects of
clinically applied drugs, the concept of “Number
Needed to Harm” (NNH) has been developed,
which is defined as the average number of people
who would need to be treated over a specific
period of time before one bad outcome of the
treatment would occur. NNH is a concept showing
how many members of a population taking a given
drug in the course of their treatment are likely to
suffer from side effects. The following example will
explain the concept. Out of 100 patients taking
drug A, 30 show weight gain as a side effect, while
of 100 patients using drug B, only 10 develop
weight gain; in drug A, the undesired result affects
30%, in drug B 10%, leading to an NNH value of 1/
(0.30-0.10)=5. That is to say, in order to
Klinik Psikofarmakoloji Bülteni, Cilt: 25, Sayı: 1, 2015 / Bulletin of Clinical Psychopharmacology, Vol: 25, N.: 1, 2015 - www.psikofarmakoloji.org
1
Number needed to treat: an important measure for the correct assessment of clinical significance
demonstrate the side effect of weight gain to be
higher in drug A than in drug B or vice versa, to
show the effect to be lower in drug B than in drug
A, at least 5 patients need to be treated. In that
case, from the perspective of the side effect of
weight gain, drug B is superior to drug A. As
explained above, the desired outcome is an NNT
value close to 1, showing the drug’s efficacy and
effectiveness, and a very large NNH value, heading
towards infinity, showing that the side effects of
the drug are very small. As a concrete example, in
a meta-analysis conducted by Stewart et al. to
assess the usefulness of antidepressants in cases of
non-severe major depression, the authors
reported that an NNT between 3 and 8 was
sufficient evidence to recommend the studied
antidepressants for therapeutic use15. In another
study assessing the efficacy of some
antipsychotics used in schizophrenic patients to
prevent relapse, NNT values between 2 and 5 have
been reported as an indicator for a medium and
broad effect size16.
A review made by Citrome mentions one
study comparing asenapine against a placebo in
schizophrenic patients. With asenapine (10mg
twice daily), an NNH of 20 was found for oral
hypoesthesia; with a dose of 5mg asenapine
twice daily, the NHH was 20 for hypoesthesia in
the mouth and 13 for somnolence. In patients
suffering from bipolar disorder, compared to
placebo, asenapine at 5-10mg twice daily
resulted in NNH values of 6 for somnolence, 13
for dizziness, 20 for extra-pyramidal symptoms
other than akathisia, and 25 for increased
weight14.
In a paper published by Srivastava and
Ketter, large-scale, randomized, double-blind,
placebo-controlled studies in acute mania and
bipolar depression were assessed with regard to
the NNT for response to treatment and NNH for
sedation and weight gain. They found that
lithium, in comparison with other FDAapproved drugs, showed a similar efficacy (NNT
4 vs. 4-8) but a significantly lower sedation
(NNH 27 vs. 5-17).
A recently published study by Ketter et al.
2
assessed old and new anti-depressants, using
mu l t i - c e n t e r, r a n d o m i z e d , d o u bl e - bl i n d ,
placebo-controlled studies and meta-analyses
using NNT values for response to treatment and
NNH for side effects. There are mainly two FDAapproved therapies for bipolar depression,
olanzapine/ fluoxetine combination (OFC) and
quetiapine monotherapy (QTP). For the response
efficacy, OFC resulted in NNT=4 and QTP NNT=6,
while side effects of weight gain (OFC: NNH=6)
and sedation/somnolence (QTP: NNH=5) were
noted. In another example, the antipsychotic
lurasidone, recently approved in the US, showed
a response NNT=5 for monotherapy and NNT=7
as adjunctive, with NNH=16 for nausea
(adjunctive) and NNH=15 for akathisia (as
monotherapy), thus being effective as well as
better tolerated17.
Which of these values do we take into
consideration, and are there optimal values for
each of these measures? The US Food and Drug
Administration (FDA) has approved drugs for the
pharmacotherapy of bipolar disorder with NNT
values in the single figures, below ten (with an
advantage over placebo of more than 10%), and
there are studies reporting that the clinical
efficacy of drugs with NNT values below ten have
been accepted as significant11,12,19,20. While there is
a greater consensus in the studies regarding
acceptable NNT values, it has been reported that
the acceptability of NNH values varies particularly
depending on the outcome of undesirable side
effects. Therefore, in more acute severe illness
presentations, drugs with high therapeutic
success may be chosen even though their side
effects are greater, while in chronic, mild to
medium cases even less efficacious drugs with
fewer side effects may be preferred. In the light of
this information, it is therefore suggested that
clinicians, when deciding about the choice of
therapy, take both efficacy and tolerability into
consideration, looking not only at the NNT values
regarding the effect, but also at the NNH to
estimate the occurrence of various side effects,
thereby making careful, individualized, patientcentered decisions.
Klinik Psikofarmakoloji Bülteni, Cilt: 25, Sayı: 1, 2015 / Bulletin of Clinical Psychopharmacology, Vol: 25, N.: 1, 2015 - www.psikofarmakoloji.org
Cetin M, Kilic S
1
Correspondence Address: Prof. Dr. Mesut Cetin,
Klinik Psikofarmakoloji Bülteni-Bulletin of Clinical
M.D., Professor of Psychiatry, Editor -in- Chief,
GATA Haydarpasa Training Hospital, Department of
Pychopharmacology, GATA Haydarpasa Training
Psychiatry, Uskudar, 34668, Istanbul-Turkey
Hospital, Department of Psychiatry, Istanbul-Turkey
Phone: +90-216-464-2888
Email address: [email protected]
2
Prof., M.D., Department of Epidemiology, Gulhane
School of Medicine, Ankara-Turkey
References:
1. Laupacis A, Sackett DL, Roberts RS. An assessment of
clinically useful measures of the consequences of
treatment. N Engl J Med 1988;318:1728-33. [CrossRef]
2. Srivastava S, Ketter TA. Clinical relevance of treatments for
acute bipolar disorder: balancing therapeutic and adverse
effects. Clin Ther 2011;33(12):B40-8. [CrossRef]
3. Cetin M, Acikel CH. In the perspective of meta-analyses: are
all of the antidepressants similar? Klinik Psikofarmakoloji
Bulteni - Bulletin of Clinical Psychopharmacology
2009;19(2):87-92. (Turkish)
4. Cetin M, Kilic S. How to comprehend the recent metaanalyses conducted on typical and atypical antipsychotics?
Klinik Psikofarmakoloji Bulteni - Bulletin of Clinical
Psychopharmacology 2009;19:(1)1-4. (Turkish)
5. Alvarez-Jiménez M, Parker AG, Hetrick SE, McGorry
PD, Gleeson JF. Preventing the second episode: a
systematic review and meta-analysis of psychosocial and
pharmacological trials in first-episode psychosis. Schizophr
Bull 2011;37(3):619-30. [CrossRef]
6. Smith EG, Søndergård L, Lopez AG, Andersen PK,
Kessing LV. Association between consistent purchase of
anticonvulsants or lithium and suicide risk: a longitudinal
cohort study from Denmark, 1995-2001. J Affect Disord
2009;117(3):162-7. [CrossRef]
7. Gao K, Ganocy SJ, Gajwani P, Muzina DJ, Kemp DE,
Calabrese JR.A review of sensitivity and tolerability
of antipsychotics in patients with bipolar disorder or
schizophrenia: focus on somnolence. J Clin Psychiatry
2008;69(2):302-9. [CrossRef]
8. Correll CU, Kishimoto T, Nielsen J, Kane JM. Quantifying
clinical relevance in the treatment of schizophrenia. Clin
Ther 2011;33(12):B16-39. [CrossRef]
9. Citrome L. Relative vs. absolute measures of benefit
and risk: what’s the difference? Acta Psychiatr Scand
2010;121(2):94-102. [CrossRef]
10. Citrome L. The Tyranny of the P-value: effect size matters.
Klinik Psikofarmakoloji Bulteni - Bulletin of Clinical
Psychopharmacology 2011;21(2):91-2. [CrossRef]
11.Citrome L, Ketter TA. When does a difference make a
difference? Interpretation of number needed to treat,
number needed to harm, and likelihood to be helped or
harmed. Int J Clin Pract 2013;67(5):407-11. [CrossRef]
12.Citrome L. Adjunctive aripiprazole, olanzapine, or
quetiapine for major depressive disorder: an analysis
of number needed to treat, number needed to harm,
and likelihood to be helped or harmed. Postgrad Med
2010;122(4):39-48. [CrossRef]
13. Correll CU, Sheridan EM, DelBello MP. Antipsychotic and
mood stabilizer efficacy and tolerability in pediatric and
adult patients with bipolar I mania: a comparative analysis
of acute, randomized, placebo-controlled trials. Bipolar
Disord 2010;12(2):116-41. [CrossRef]
14.Citrome L. Asenapine for schizophrenia and bipolar
disorder: a review of the efficacy and safety profile for this
newly approved sublingually absorbed second-generation
antipsychotic. Int J Clin Pract 2009;63(12):1762-84. [CrossRef]
15.Stewart JA, Deliyannides DA, Hellerstein DJ, McGrath PJ,
Stewart JW. Can people with nonsevere major depression
benefit from antidepressant medication? J Clin Psychiatry
2012;73(4):518-25. [CrossRef]
16. Gopal S, Berwaerts J, Nuamah I, Akhras K, Coppola D, Daly E,
Hough D, Palumbo J. Number needed to treat and number
needed to harm with paliperidone palmitate relative to
longacting haloperidol, bromperidol, and fluphenazine
decanoate for treatment of patients with schizophrenia.
Neuropsychiatr Dis Treat 2011;7:93-101. [CrossRef]
17.Ketter TA, Miller S, Dell’Osso B, Calabrese JR, Frye MA,
Citrome L. Balancing benefits and harms of treatments for
acute bipolar depression. J Affect Disord 2014;169(Suppl
1):S24-S33. [CrossRef]
18.Citrome L. Lurasidone for the acute treatment of adults
with schizophrenia: what is the number needed to treat,
number needed to harm, and likelihood to be helped or
harmed? Clin Schizophr Relat Psychoses 2012;6(2):76-85.
[CrossRef]
19.Ketter TA, Citrome L, Wang PW, Culver JL, Srivastava S.
Treatments for bipolar disorder: can number needed to
treat/harm help inform clinical decisions? Acta Psychiatr
Scand 2011;123(3):175-89. [CrossRef]
20.Popovic D, Reinares M, Amann B, Salamero M, Vieta
E. Number needed to treat analyses of drugs used
for maintenance treatment of bipolar disorder.
Psychopharmacology (Berl) 2011;213:657-67. [CrossRef]
Klinik Psikofarmakoloji Bülteni, Cilt: 25, Sayı: 1, 2015 / Bulletin of Clinical Psychopharmacology, Vol: 25, N.: 1, 2015 - www.psikofarmakoloji.org
3