NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Wilt TJ, Shaukat A, Shamliyan T, et al. Lactose Intolerance and Health. Rockville (MD): Agency for Healthcare Research and Quality (US); 2010 Feb. (Evidence Reports/Technology Assessments, No. 192.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of Lactose Intolerance and Health

Lactose Intolerance and Health.

Show details



Analytic Framework

We followed the analytic framework (modified from the U.S. Preventive Services Task Force)8 to determine causality between treatments and patient outcomes and adverse events in patient subpopulations, including age, race, and ethnic subgroups. Probabilities of diagnosis, treatment, and outcomes were analyzed based on the published literature.

Figure 1 describes target population and also includes individuals with self reported LI (regardless of symptoms) as well as individuals with clinically diagnosed LI, which may include those with lactose malabsorption, lactase nonpersistence, etc. Figure 1 also gives information about research questions:

Figure 1. Analytic framework.

Figure 1

Analytic framework.

  • KQ1. What is the prevalence of lactose intolerance?
  • KQ2. What are the intermediate and clinical outcomes of lactose free or low lactose diets?
  • KQ3. What amount of daily lactose intake is tolerable in subjects with diagnosed lactose intolerance?
  • KQ4. What are the intermediate, clinical, and adverse outcomes after treatments for lactose intolerance?

In the clinical situation, a graduated definition of a potentially lactose intolerant subject, might be as follows:

  1. The quantity of lactose routinely ingested by the individual that causes symptoms.
  2. The quantity of lactose ingested in some situations by the individual causes the above symptoms.
  3. The quantity of lactose that the individual would like to ingest (but does not due to fear of symptoms) causes the above symptoms.
  4. The quantity of lactose ingested in the course of obtaining 1,500 mg/day of calcium entirely via lactose-containing dairy products causes the above symptoms.

A confounding problem is that factors other than simply the quantity of lactose ingested might influence a subject’s symptomatic response, i.e., the form in which lactose is ingested (ice cream versus milk, etc.), the coingestion of nonlactose containing foods, the nonspecificity of symptoms, and the large placebo response potentially observed.

Criteria for Inclusion/Exclusion of Studies in Reviewing and Searching for the Evidence: Literature Search Strategies for Identification of Relevant Studies to Answer the Key Questions

General Inclusion Criteria

We included original observational studies that examined prevalence, symptoms, and outcomes in different age, gender, racial, and ethnic groups; published in the English language; randomized controlled clinical trials (RCTs) that examined different treatment options or doses of lactose loads in patients with LI or LM; and large observational studies in individuals with LI, LM, lactase nonpersistence, or reduced dairy intake that performed at least one strategy to reduce bias. We limited our search to studies published from 1967 to November 2009. We excluded studies that were published in non English languages and small case reports or descriptive case series with less than 100 subjects unless there are no reliable data from other higher quality studies. Because this report is to be used for a U.S. NIH Consensus Conference report we emphasized U.S. based population studies. We excluded populations with other GI disorders, including individuals diagnosed with IBS, inflammatory or infectious bowel diseases, or milk allergies. We excluded children younger than 4 years of age.

Key Question 1: What is the prevalence of lactose intolerance? How does this differ by race, ethnicity, and age?

Study eligibility. We included studies if: (1) they were original research articles, (2) they presented prevalence data related to nonacute LI or LM, including self-reported symptoms, symptoms following a lactose challenge, symptoms following a placebo controlled and blinded lactose challenge, lactose malabsorption via a hydrogen breath test following a lactose challenge, hypolactasia defined by biopsy or genetic tests for adult-type hypolactasia, (3) the study population was not primarily secondary lactose intolerance related to other conditions or treatments, and (4) only results for those greater than 1 year of age. Since the focus of this report is to provide evidence most relevant for a U.S. population, all studies with a sample size greater than 50 that met the previous criteria were included if the study reported results from a U.S. population. Only larger studies (at least 100 participants) of populations outside of the United States were included.

To the extent that evidence of reliable estimates of LI is missing, we reviewed the evidence of prevalence of lactose malabsorption, lactase nonpersistence (adult-type hypolactasia) and self-reported symptoms following lactose consumption.

Population. We included persons older than 4 years of age.

Conditions. We defined lactose intolerance to be present when ingestion of 50 grams of lactose (or less) as a single dose by a lactose malabsorbing subject induces GI symptoms (abdominal pain, bloating, diarrhea, nausea, flatulence) not observed when the subject ingests an indistinguishable placebo. The 50 gram dose of lactose, the quantity present in a quart of milk, was selected because this quantity of milk provides the maximal recommended daily intake of calcium (1,500 mg), and this dosage approaches the maximal daily volume of milk likely to be ingested by most Americans ( As discussed previously, LI defined in this way does not indicate that intolerance symptoms necessarily will be recognizable when these subjects ingest smaller dosages of lactose (as does the vast majority of the U.S. population). The prevalence of clinically important LI requires demonstration that the quantity of lactose that subjects actually ingest (or wish to ingest) causes symptoms in placebo-controlled experiments. We excluded congenital lactase deficiency, developmental lactase deficiency among pre-term infants, milk allergies commonly seen in infants, and acute lactose intolerance (<30–60 days duration) due to such things as antibiotic use or illness.

Disease severity. Lactose malabsorption is the physiologic problem that manifests as LI and is attributable to an imbalance between the amount of ingested lactose and the capacity for lactase to hydrolyze the disaccharide.7 LM indicates that a sizable fraction of a dosage of lactose is not absorbed in the small bowel and thus is delivered to the colon. We defined severity of LM according to the amount of consumed lactose (desired or required to meet established dietary needs) before experiencing clinical symptoms of LI. Since such malabsorption virtually always is a result of low levels of lactase, there is a nearly one to one relationship of lactase nonpersistence (or deficiency) and LM. LM is objectively demonstrated via measurements of breath H2 or blood glucose concentrations following ingestion of a lactose load. We analyzed severity of lactose intolerance according to criteria from diagnostic tests: lactose intolerance breath test: increase from baseline in hydrogen + methane (in parts per million [ppm]) by 20–38 ppm as mild and >39 as severe LI.

We defined lactase nonpersistence according to presence of lactase enzyme on intestinal biopsy and according to the presence of the C/C genotype of the lactase promoter gene with genetic testing using restriction fragment length polymorphism or by DNA Sequencing to detect single-nucleotide polymorphisms (C-13910T, G-22018A) located upstream of the lactase gene within the gene MCM6.

We reviewed differences in prevalence estimates based on different definitions of LI:

  • Primary lactase deficiency, a genetically determined decrease or absence of lactase is noted, while all other aspects of both intestinal absorption and brush border enzymes are normal. Primary lactase deficiency is attributable to relative or absolute absence of lactase that develops in childhood at various ages in different racial groups and is the most common cause of LM and LI. Primary lactase deficiency is also referred to as adult-type hypolactasia, lactase nonpersistence, or hereditary lactase deficiency.
  • We excluded individuals if secondary lactase deficiency occurs in association with small intestinal mucosal disease with abnormalities in both structure and function of other brush border enzymes and transport processes. Secondary lactase deficiency is often seen in celiac sprue.

Comorbidities, patient demographics. We attempted to review differences in prevalence in individuals of different age groups defined as: Preschool Children: 4–5 years, Children: 6–12 years, Adolescents: 13–18 years, Adults: 19–44 years, Middle Aged: 45–64 years, Aged: 65+ years, and Elderly Adults: 80 and over.

We attempted to review differences in prevalence of LI in patients of different race-ethnicity groups defined as: Continental Africans, Asians and Europeans, African Americans, Arabs, Caucasians, Arabs, Asian Americans, and Hispanic Americans. We included studies of patients with LI and all comorbidities except acute diseases, treatment of which could cause secondary LI.


Prevalence of LI. We reported prevalence according to: (a) patient reported diagnosis of LI, (b) clinician diagnosis of LI, and (c) absolute difference in prevalence of individuals with symptoms as derived from randomized controlled blinded trials conducted in subjects diagnosed with LI. We compared outcomes between individuals with a diagnosis of LI receiving blinded lactose (at varying doses) and control interventions, as well as the outcome from blinded RCTs, comparing outcomes in subjects with diagnosed LI versus control subjects. We assessed the prevalence of LM by evaluating studies using breath hydrogen measures.

Glucose tolerance testing is rarely used clinically today, and studies assessing this method for evaluating LM were excluded. Studies assessing only intestinal biopsies were reviewed for quality and applicability.

A critical aspect of this question was to clearly define and differentiate between: (1) lactase nonpersisters, (2) lactose malabsorbers, and (3) lactose intolerance.

LI is the key component of this question and conference. Identifying a gold-standard definition of LI is critical and difficult. There is no objective laboratory test (intestinal biopsies are rarely done and only assess lactase enzyme levels; physiologic tests: e.g., hydrogen breath tests measure LM to a laboratory challenge and need to be evaluated to determine whether they accurately identify clinically relevant LI.

We defined LI to be present when ingestion of 50 grams of lactose (or less) as a single dose by a lactose malabsorbing subject induces gastrointestinal symptoms not observed when the subject ingests an indistinguishable placebo.

We evaluated prevalence according to different populations and methods of assessment with a particular focus on presence or absence of specific symptoms among individuals participating in blinded RCTs evaluating LI. While assessing prevalence in RCTs typically is not done to assess prevalence, we believe that patient reported symptoms and resolution of symptoms in the absence of placebo controlled trials are not reliable.

Key Question 2: What are the health outcomes of dairy exclusion diets?

Population. We included populations that consumed or were likely to consume dairy free or low dairy diets and reported on long-term GI and bone outcomes. We excluded individuals with irritable bowel syndrome or other GI disorders, such as infectious or inflammatory diarrhea. We excluded populations with children under age 4.

Interventions. We defined dairy exclusion diets as low lactose diets that generally eliminate only milk and milk products or lactose free diets that eliminate all lactose products, including foods that are prepared with milk, both at home and in commercially packaged foods. We included studies with the following comparators: placebo or regular diet. We defined interventions when patients followed lactose free diets prescribed by health care professionals. We defined exposure when subjects followed low lactose or lactose free diets without recommendations from health care professionals. We included indirect evidence of the effect of dairy exclusion on health outcomes by including studies of populations known or suspected of having low dairy intake (e.g., diagnosis of LI/LM, lactase nonpersistence based on intestinal biopsy or genetic test association for lactase nonpersistence) even in the absence of specific documentation of amount of lactose intake. We assessed associations between lactose intake and factors associated with low lactose intake on GI symptoms or bone health, including clinical (fracture) and intermediate outcomes (osteoporosis, bone mineral density, and content).


Primary bone outcomes. Fracture.

Secondary bone outcomes. Osteoporosis, bone density, bone content.

Primary gastrointestinal outcomes. Abdominal pain, diarrhea, nausea, flatulence, bloating. Osteoporosis was defined according to World Health Organization Criteria1–3 as a BMD 2.5 standard deviation or more below the young average value in women and men.4 Osteopenia was defined as a BMD 1–2.5 standard deviation below the population average.5

We used reference data on femur bone mineral content and density of noninstitutionalized adults in the United States from the third National Health and Nutrition Examination Survey that collected dual energy x-ray absorptiometry in the nationally representative sample of 14,646 men and women 20 years of age and older.6

Adverse events. All published adverse events.

Timing. We included prospective and retrospective studies with duration of followup long enough to detect long-term differences in outcomes (5 years for fractures, 1–2 years for secondary bone outcomes, and greater than 1 month for GI symptoms). We evaluated the impact of lactose exclusion diets on shorter-term (<1 month) patient reported GI symptoms from observational and interventional studies among individuals with both LI and non LI controls. GI outcomes from RCTs with shorter duration followup are reported in Key Questions 3 and 4.

Setting. We included studies in primary and specialty outpatient settings and population based settings.

Co-interventions. We reviewed co-interventions in studies that reported patient outcomes after low lactose and lactose free diets.

We conducted a literature search to identify three types of studies:

  1. Studies in patients with LI who followed lactose free diets.
  2. Studies that examined patient outcomes among healthy populations consuming dairy exclusion (or very low dairy) diets (e.g., vegans).
  3. Meta-analyses and systematic reviews that synthesized the association between dairy (dietary Ca++) intake and patient outcomes.

Confounding factors. We analyzed the adjustment for the known factors that could confound the association between lactose intake and bone health, including age, gender, race, menopausal status in women, external calcium supplementation, renal function, and smoking. We abstracted how systematic reviews addressed the adjustment for confounding for the association between low milk intake and bone fractures.

The main long-term health concern related to lactose exclusion diets from this report was predominately related to potentially low calcium and vitamin D intake associated with these diets. We also assessed the impact of dietary or supplemental calcium and/or vitamin D. We reviewed whether the studies that examined patient outcomes in association with low dietary milk intake addressed calcium intake from other sources and supplementation with Ca++ or vitamin D. This provided us with contextual information regarding the potential role of low lactose or lactose free diet on bone health independent of other sources of Ca++ or vitamin D.

Key Question 3: What amount of daily lactose intake is tolerable in subjects with diagnosed lactose intolerance?

Population. Our target population was limited to subjects with self or clinically diagnosed LI. We focused on populations with clinically diagnosed LI. We defined genetic testing as reference methods to diagnose primary LI. We defined LI breath tests as methods for assessing LM. We defined self reported LI as the presence of self described clinical symptoms occurring only when they ingested lactose and relieved when they eliminated/reduced lactose or used products to hydrolyze lactose prior to ingestion. We quantified the type and severity of symptoms and the amount and type of lactose causing patient reported symptoms. A presumptive working diagnosis of LI was GI symptoms associated with the ingestion of foods containing a quantity of lactose that is either desired by the individual or considered necessary to meet national minimal daily dietary standards, and that resolve upon elimination or marked reduction of these lactose containing foods or when using products to hydrolyze lactose prior to ingestion and return upon lactose rechallenge provided in a blinded fashion. We defined self reporting as index methods to diagnose LI.

Interventions. We evaluated individual daily or weekly intake of lactose stratified by the presence or absence of index diagnostic tests for LI.

Comparators. Placebo, inactive comparator, lactose dose response.


Primary outcomes. Our primary outcomes included the prevalence and severity of GI symptoms, particularly abdominal pain, diarrhea, nausea, flatulence, and/or bloating. We assessed for the percentage reporting these outcomes as well as scores reported on symptom questionnaires.

Timing. Short term (≤1 month) long-term (>6 months).

Settings. Primary and specialty outpatient settings, population based settings.

Because there was strong evidence of a placebo response, we relied on an evaluation of results from blinded RCTs, including dosing studies to determine the threshold amounts that caused symptoms in subjects with self or clinician diagnosed LI (with or without laboratory evidence of LM) ingesting different doses of lactose versus controls and the outcomes among individuals ingesting lactose with a diagnosis of LI versus non LI controls. Where possible, we attempted to categorize findings according to age, ethnicity, and patient reported baseline LI severity and whether symptoms differed between subjects diagnosed with LI (self versus clinician) and controls.

We characterized the diagnostic standards used in these studies (e.g., patient reported symptoms and breath hydrogen (measure of LM not LI). If there are gaps in evidence related to amount and type of daily lactose intake, symptoms were defined as patient reported: gas/flatulence, abdominal pain, bloating, and diarrhea.

Key Question 4: What strategies are effective in managing individuals with diagnosed lactose intolerance?

Study inclusion. We included randomized double blind controlled trials that evaluated probiotics, enzyme replacement therapies with lactase from nonhuman sources, administration of lactose reduced milk, and regimes of increases in dietary lactose load for improvement of GI symptoms in individuals with presumed LI or LM.

Population. Subjects with presumed LI, LM, or controls and greater than 4 years of age. We also included double blind randomized trials that enrolled subjects with IBS and LM or LI. These were reported as a separate group. We excluded individuals with presumed IBS alone and other likely causes of acute GI symptoms (e.g., infectious, antibiotic, or inflammatory associated bowel disease).

Interventions. We evaluated the following interventions:

  • Commercially available lactase
  • Prebiotics and probiotics
  • Incremental lactose loads for colonic adaptation
  • Other dietary strategies

Comparators. Placebo, usual care, no active treatment, or active control.


Primary outcomes. Disease specific and overall quality of life.

Secondary outcomes. Frequency and severity of specific GI items of disease specific quality of life questionnaires: abdominal pain, diarrhea, nausea, flatulence, bloating.

Adverse events. We evaluated all published adverse events.

Timing. We analyzed all eligible studies regardless of followup duration.

Settings. We included primary and specialty outpatient settings and population based settings.

Assessment of Methodological Quality of Individual Studies

We rated the quality of studies according to recommendations from the Methods Guide for Comparative Effectiveness Reviews. We used the following ratings of Quality of Individual Studies:

  • Well designed and conducted (good; low risk of bias). A study that adheres mostly to the commonly held concepts of high quality, including the following: a formal randomized controlled study; clear description of the population, setting, interventions, and comparison groups; appropriate measurement of outcomes; appropriate statistical and analytic methods and reporting; no reporting errors; low dropout rate; and clear reporting of dropouts.
  • Fair. These studies are susceptible to some bias, but it is not sufficient to invalidate the results. They do not meet all the criteria required for a rating of good quality because they have some deficiencies, but no flaw is likely to cause major bias. The study may be missing information, making it difficult to assess limitations and potential problems.
  • Poor (high risk of bias). These studies have significant flaws that imply biases of various types that may invalidate the results. They have serious errors in design, analysis, or reporting; large amounts of missing information; or discrepancies in reporting.

We assessed for external validity (applicability) according to the Methods Guide for Comparative Effectiveness Reviews.

Data Synthesis

We summarized evidence into summary tables with qualitative analysis of the results for prevalence of LI by subgroups for Key Question 1. We did not pool results for Key Question 1.We attempted to calculate odds ratio with 95 percent confidence interval (CI) or absolute risk differences from the reported number of events in RCTs as well as the number needed to treat to achieve one event of the outcome if the data are homogeneous enough to permit pooling. All additional calculations were performed at 95 percent confidence levels.

We calculated minimum difference in continuous variables from the reported sample size, means, and standard deviations. We calculated crude odds ratios from the reported number of subjects with and without outcomes among compared categories of exposure. Calculations were performed using STATA software,152 SAS 9.2,153 and Meta-analyst software (available at at the 95 percent confidence level.

Attributable risk was calculated as the outcome events rate in patients exposed to different clinical interventions.9–11 The number needed to treat to prevent one symptomatic event was calculated as the reciprocal to the absolute risk differences in rates of outcomes events in the active and control groups: 1/(control group event rate - treatment group event rate).10–12 We did not pool data related to Key Questions 1 or 2.

For Key Questions 3 and 4 if symptoms associated with lactose malabsorption (abdominal pain and frequency of diarrhea) data were appropriate for pooling, they were analyzed using RevMan 5.0 software using a random effects model.154 Standardized mean differences (symptom effect sizes) were calculated with the generic inverse variance method due to the crossover study design of the trials.

Grading the Evidence for Each Key Question

Assess Study Quality and Strength of Evidence

On the basis of the quality checklist(s) developed for articles relevant to the various key questions, we assigned a quality score to each article. We used methods for assessing study quality and strength of evidence according to the Methods Guide for comparative Effectiveness Reviews that is conceptually similar to the GRADE (Grades of Recommendation Assessment, Development, and Evaluation) system of evidence rating.13,14 Specifically, we assessed four domains: risk of bias, consistency, directness, and precision. When appropriate, we also include dose response association, presence of confounders that would diminish an observed effect, strength of association, and publication.

Quality of evidence across studies for each outcome. We graded the quality of evidence for primary outcomes across studies as illustrated below:

Overall ranking of evidence

HighHigh confidence that the evidence reflects the true effect. Further research is very unlikely to change our confidence in the estimate of effect.
ModerateModerate confidence that the evidence reflects the true effect. Further research may change our confidence in the estimate of effect and may change the estimate.
LowLow confidence that the evidence reflects the true effect. Further research is likely to change the confidence in the estimate of effect and is likely to change the estimate.
InsufficientEvidence either is unavailable or does not permit a conclusion.


Appendixes and evidence tables cited in this report are available at http://www​​/pub/evidence​/pdf/lactoseint/lactint.pdf.


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...