Vitamina D para el tratamiento de las afecciones dolorosas crónicas en adultos
Resumen
Antecedentes
Esta revisión es una actualización de una revisión publicada anteriormente en la Base de Datos Cochrane de Revisiones Sistemáticas (The Cochrane Database of Systematic Reviews) (Número 1, 2010) sobre "Vitamina D para el tratamiento de afecciones dolorosas crónicas en adultos”.
La vitamina D se sintetiza en la piel después de la exposición a la luz solar y puede obtenerse a través de los alimentos. La deficiencia de vitamina D se ha vinculado con varias afecciones, incluido el dolor crónico. Pruebas observacionales y circunstanciales indican que la deficiencia de vitamina D puede estar relacionada con la etiología de las afecciones dolorosas crónicas.
Objetivos
Evaluar la eficacia y la seguridad de los suplementos de vitamina D en las afecciones dolorosas crónicas en contraposición con el placebo o comparadores activos.
Métodos de búsqueda
Para esta actualización, se hicieron búsquedas en el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials, CENTRAL), MEDLINE y EMBASE hasta febrero de 2015. Se buscaron estudios adicionales en las listas de referencias de los artículos, las revisiones del área y los registros de ensayos obtenidos.
Criterios de selección
Se incluyeron los estudios que fueran ensayos a doble ciego aleatorios de suplementos de la vitamina D en comparación con un placebo u otros comparadores activos para el tratamiento de las afecciones dolorosas crónicas en adultos.
Obtención y análisis de los datos
Dos autores de la revisión seleccionaron de forma independiente los estudios para su inclusión, evaluaron la calidad metodológica y extrajeron los datos. No se realizó un análisis agrupado debido a la heterogeneidad de los datos. Los resultados primarios de interés fueron los resultados del dolor en pacientes que respondieron al tratamiento, y los resultados secundarios fueron los resultados del dolor promedio del grupo de tratamiento y los eventos adversos.
Resultados principales
Se incluyeron seis nuevos estudios (517 participantes) para esta actualización de revisión, con lo que el total de los estudios incluidos ascendió a 10 (811 participantes). Los estudios eran heterogéneos con respecto a la calidad del estudio, las afecciones dolorosas crónicas que se investigaron, la dosis de vitamina D administrada, las cointervenciones y las medidas de resultado informadas. Sólo dos estudios informaron sobre los resultados del dolor en pacientes que respondieron; los otros estudios informaron sobre los resultados promedio del grupo de tratamiento solamente. En general, no hubo una pauta consistente en que el tratamiento con vitamina D se asociara con una mayor eficacia que el placebo en ninguna afección dolorosa crónica (evidencia de baja calidad). Los eventos adversos y los retiros fueron comparativamente infrecuentes, sin una diferencia consistente entre la vitamina D y el placebo (evidencia de buena calidad).
Conclusiones de los autores
Las pruebas que abordan el uso de la vitamina D para el dolor crónico contienen ahora más del doble de estudios y participantes que los que se incluyeron en la versión original de esta revisión. Sobre la base de estas pruebas, es poco probable que la vitamina D tenga un gran efecto beneficioso en diferentes afecciones dolorosas crónicas. Es necesario investigar más a fondo si la vitamina D puede tener efectos beneficiosos en afecciones dolorosas crónicas concretas.
PICO
Resumen en términos sencillos
Vitamina D para el tratamiento de las afecciones dolorosas crónicas en adultos
El dolor crónico es dolor de intensidad moderada o grave y dura tres meses o más. Puede tener varias causas, pero la mayoría proviene de afecciones musculoesqueléticas como la artritis o el dolor muscular. El dolor crónico suele afectar más a las personas mayores que a las jóvenes. El dolor crónico es incapacitante y tiene un gran impacto negativo en la calidad de vida.
La vitamina D cumple varias funciones en el organismo. Se produce en la piel mediante la acción de la luz solar y también puede obtenerse a partir de alimentos. Se ha implicado la falta de vitamina D con varias afecciones, incluido el dolor crónico. Además, la asociación de dolores tan diversos como el dolor de cabeza, el dolor abdominal, el dolor de rodilla y el dolor de espalda con la estación del año y la latitud, apoyan indirectamente la función de la vitamina D. La posibilidad de que exista una relación entre los bajos niveles de vitamina D y el dolor crónico ha despertado interés porque, de ser cierto, la vitamina D sería un tratamiento barato y relativamente seguro.
Se buscaron en las bases de datos científicas estudios que compararan la suplementación con vitamina D con placebo (un tratamiento falso) o medicamentos activos para el tratamiento de afecciones dolorosas crónicas en adultos. La evidencia está actualizada hasta febrero de 2015.
Hay una pequeña cantidad de pruebas que sostienen esta relación pero no es de alta calidad y podría no ser fiable. Esta actualización de una revisión buscó evidencia de alta calidad de ensayos controlados aleatorios (estudios en los que los participantes se asignan al azar para recibir uno de varios tratamientos) sobre la vitamina D para las afecciones dolorosas crónicas.
No se encontró un patrón consistente de que el tratamiento con vitamina D fuera mejor que el placebo para cualquier condición dolorosa crónica, pero los estudios tenían deficiencias metodológicas (evidencia de baja calidad).
Se necesitan más estudios para determinar si la vitamina D es un tratamiento útil para el dolor en alguna afección dolorosa crónica en particular. Estos estudios deberían examinar si algún efecto se limita a las personas con deficiencia de vitamina D. También deberían examinar cuánta vitamina D es necesaria, y durante cuánto tiempo, antes de que se produzcan efectos beneficiosos.
Authors' conclusions
Background
This review is an update of a previously published review in the Cochrane Database of Systematic Reviews (Issue 1, 2010) on 'Vitamin D for the treatment of chronic painful conditions in adults'.
Vitamin D (calciferol) is a fat‐soluble vitamin that is synthesised from a precursor in the skin after exposure to ultraviolet B (UVB) radiation from the sun, and it is also obtained from dietary sources (eg oily fish, vitamin D‐fortified milk and breakfast cereals, or dietary supplements). Vitamin D can be of plant origin (vitamin D2 or ergocalciferol) or of animal origin (vitamin D3 or cholecalciferol). Vitamin D from the skin and diet is metabolised in the liver to 25‐hydroxyvitamin D, and further metabolised in the kidneys to its active form, 1,25‐dihydroxyvitamin D. Vitamin D exerts its effects by modulating gene expression after binding to a nuclear vitamin D receptor. Most tissues in the human body express vitamin D receptors (Holick 2007), supporting the concept of a wide‐ranging importance of vitamin D.
Vitamin D deficiency can arise from a number of causes including insufficient UVB exposure, decreased bioavailability (eg malabsorption disorders), or diseases that affect the metabolism of vitamin D, such as liver and kidney disease. Some medications can also cause vitamin D deficiency. These include anticonvulsants, glucocorticoids, and antiretroviral drugs (Holick 2007).
Lack of vitamin D is known to be implicated in diseases of the musculoskeletal system, where its deficiency is classically a cause of rickets (in children) and osteomalacia (in adults). Insufficient vitamin D has been implicated in a range of diseases involving other body systems, including metabolic, autoimmune, psychiatric, respiratory, and cardiovascular disorders; cancers (especially colon, prostate, and breast cancer); and chronic pain (Gröber 2013; Holick 2007). It has been suggested that there is widespread vitamin D deficiency or insufficiency in many populations and that sensible sun exposure and vitamin D supplementation may be called for (Gröber 2013; Holick 2007). Indeed, one meta‐analysis found a reduced all‐cause mortality with vitamin D supplementation (Autier 2007). One more recent Cochrane review found that, overall, vitamin D slightly decreased mortality, but further analyses suggested that the benefit in terms of vitamin D supplementation decreasing mortality was more limited; there may be benefit with vitamin D3 supplementation for elderly people (Bjelakovic 2014).
The Institute of Medicine recommends a dietary intake of 600 international units (IU) of vitamin D (800 IU for people older than 70 years) (IOM 2011), other publications recommend higher intakes, at least without adequate sun exposure (Gröber 2013; Holick 2007). One IU corresponds to 25 ng of vitamin D.
The precise definition of vitamin D deficiency is a matter of debate. It has been argued that serum 25‐hydroxyvitamin D levels below 20 ng/mL signify pronounced deficiency and that levels between 21 and 29 ng/mL signify moderate deficiency, or insufficiency (Gröber 2013; Hossein‐nezhad 2013). It has further been suggested that a 25‐hydroxyvitamin D level between 40 and 60 ng/mL is ideal and that toxicity is expected only at levels greater than 150 ng/mL (Gröber 2013). For the purpose of this review, we have regarded levels of vitamin D below 20 ng/mL as evidence of deficiency, levels between 20 and 29 ng/mL as evidence of insufficiency, and levels of 30 ng/mL or greater as evidence of sufficiency.
Description of the condition
Chronic pain is generally described as pain experienced on most days for at least three months. Chronic pain in its various forms is very common, one of the most common health problems of all: one recent review of prevalence studies estimated the average prevalence of non‐cancer chronic pain in adults at about 20% (Moore 2014), with some individual studies claiming substantially higher prevalence rates. The prevalence of benign chronic back pain alone has been estimated to be between 2% and 40%, depending on the population studied (Verhaak 1998). The prevalence of arthritis was estimated at almost 16% in one large multinational study (Alonso 2004). Fibromyalgia is another chronic pain disorder that affects 1% to 6% of the population (Bannwarth 2009; Mas 2008; McQuay 2007; Vincent 2013). Furthermore, acute pain can become chronic. For example, one systematic review showed that chronic pain after surgery was surprisingly common, up to 47% after thoracic surgery (Perkins 2000). Not surprisingly, chronic pain conditions can have a profound effect on functioning and quality of life (Alonso 2004; Mas 2008; Moore 2014), with considerable economic and social impact. Chronic pain is a common cause of sickness absence from work (Moore 2014; Saastamoinen 2009).
Low levels of 25‐hydroxyvitamin D have been linked with a higher incidence of chronic pain (Atherton 2009; Benson 2006; Lotfi 2007). Additionally, associations of such diverse types of pain as headache, abdominal pain, knee pain, and back pain with season of the year and latitude may suggest that 25‐hydroxyvitamin D levels are important in this context (Mitsikostas 1996; Saps 2008; Zeng 2004). Therefore, it seems possible that vitamin D deficiency may be involved in the aetiology of chronic pain.
Description of the intervention
Vitamin D supplementation can be given orally or parenterally. A number of different preparations and dosing regimens of vitamin D treatment have been employed in clinical practice. This review considered all vitamin D preparations and dosing regimens when used in the treatment of chronic pain. Treatment with vitamin D holds great appeal because it is inexpensive; has relatively few, usually mild, adverse effects; and there is a positive public perception of vitamin supplementation that could result in high rates of adherence. Excessive vitamin D intake can result in hypercalcaemia, but this is rare.
How the intervention might work
Vitamin D supplementation increases 25‐hydroxyvitamin D blood levels (25‐hydroxyvitamin D is the form of the vitamin usually measured and recommended to be measured (Holick 2011), with assays normally detecting vitamin D2 and vitamin D3) and can, therefore, potentially correct the effects of vitamin D deficiency (Gröber 2013; Holick 2007). The details of how this might work (molecular mechanism, time to effect, extent of reversibility of pain potentially associated with vitamin D deficiency) are unclear at present.
Why it is important to do this review
At the time that the original version of this Cochrane review was written, no systematic review of vitamin D supplementation in chronic pain had been published in the Cochrane Database of Systematic Reviews. A systematic review of vitamin D treatment for chronic pain was necessary to answer the questions of whether chronic pain could be helped by vitamin D supplementation and, indirectly, whether low levels of 25‐hydroxyvitamin D in people with chronic pain were causal or coincidental.
Our initial review on this topic was published in "Pain". This was based on a MEDLINE (PubMed) search and used evidence from a variety of study architectures (Straube 2009). We felt it was important to develop this further by using a more extensive search strategy and, at the same time, more stringent inclusion criteria to focus on the high quality evidence that is most relevant to clinical practice for publication in The Cochrane Library. The original version of this Cochrane review was published in 2010 (Straube 2010). A number of studies investigating vitamin D for chronic painful conditions have been published since that time, so that an update of our Cochrane review was called for.
Objectives
To assess the efficacy and safety of vitamin D supplementation in chronic painful conditions when tested against placebo or against active comparators.
Methods
Criteria for considering studies for this review
Types of studies
We included studies if they were randomised, double‐blind, controlled trials of vitamin D supplementation compared with placebo or with active interventions in the treatment of chronic painful conditions. Randomisation and double‐blinding are known to minimise bias and were therefore deemed essential study features (Moore 2006; Moore 2010). Parallel‐group design and cross‐over studies were both acceptable.
We excluded the following types of report.
-
Review articles, case series, case reports, and clinical observations.
-
Studies of experimentally induced pain.
-
Studies where pain relief (or pain intensity with baseline level) was not measured or not patient‐reported.
Types of participants
Studies of adults (generally over 15 years of age) with all types of chronic pain conditions were eligible for inclusion. To qualify for inclusion, all study participants needed to have a chronic painful condition. Studies in conditions that were painful in some people but could be asymptomatic in other people (eg osteoporosis) did not qualify for inclusion if they did not explicitly state that all study participants had chronic pain.
Types of interventions
We compared vitamin D supplementation (oral or parenteral) with placebo or active comparator, with no restriction regarding type, dose, and frequency of vitamin D treatment. We included studies that investigated combined treatment with vitamin D and other concurrent interventions only if the other interventions were also given to the comparator group, because we wanted to isolate the treatment effect due to vitamin D.
Types of outcome measures
We collected data on participant characteristics (age, sex, and condition to be treated) and on baseline and end‐of‐trial 25‐hydroxyvitamin D levels where they were reported.
Primary outcomes
The primary outcome was patient‐reported clinically significant pain relief. We used the following hierarchy of outcome measures, in order of preference.
-
Number of participants with at least 50% pain relief.
-
Number of participants reporting a global impression of change of "much" or "very much" improvement.
-
Number of participants with undefined (any) improvement in pain.
Secondary outcomes
-
Other patient‐rated pain outcomes (typically reported as group mean values).
-
Numbers of participants with adverse events: any adverse event, any serious adverse event.
-
Numbers of withdrawals: all‐cause, lack of efficacy, and adverse event withdrawals.
-
Quality of life (any scale).
Search methods for identification of studies
We identified studies primarily by searching electronic databases, with no language restriction.
Electronic searches
For the original version of this review, the following electronic databases were searched, as far back as they would allow, without specified start dates: Cochrane Central Register of Controlled Trials (CENTRAL; Issue 3, 2009), MEDLINE (to September 2009), EMBASE (to September 2009), and the Oxford Pain Relief Database (Jadad 1996a).
For this update of the review, the trials search co‐ordinator of the Cochrane Pain, Palliative and Supportive Care group and the review authors conducted searches in the databases listed below.
-
CENTRAL (via CRSO, 2009 to 2 February 2015).
-
MEDLINE (via Ovid, September 2009 to 2 February 2015).
-
EMBASE (via Ovid, September 2009 to 2 February 2015).
See Appendix 1 for the CENTRAL search strategy, Appendix 2 for the MEDLINE search strategy, and Appendix 3 for the EMBASE search strategy.
We also searched MEDLINE (via PubMed) in December 2014 using a number of search terms to identify chronic diseases in combination with terms to identify vitamin D treatment, and we searched using the "related citations" feature in PubMed for each of the included studies.
The Oxford Pain Relief Database is no longer being updated and was not searched again for this update.
Searching other resources
We searched additional studies for this update from the reference lists of the retrieved articles and reviews in the field. We searched two online trial registries for trials that have not yet been published or are still ongoing: ClinicalTrial.gov (clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trial Registry Platform (ICTRP) (apps.who.int/trialsearch/). These trial registries were searched on 3 December 2014. We did not contact manufacturers of vitamin D preparations since vitamin D does not have an indication in chronic pain.
Data collection and analysis
Review authors were not blinded to the study authors' names and institutions, journal of publication, or study results at any stage of the review. Two review authors independently selected the studies identified for this review update for inclusion, assessed methodological study quality, and extracted data. We resolved disagreements through discussion.
Selection of studies
We assessed titles and abstracts of publications identified by the searches on screen to eliminate studies that obviously did not satisfy the inclusion criteria. We obtained full publications of the remaining studies to determine inclusion in the review.
Data extraction and management
Two review authors independently extracted data using a standard form.
Assessment of risk of bias in included studies
We used the Oxford Quality Scale score as the basis for inclusion, limiting inclusion to studies that were randomised and double‐blind as a minimum (Jadad 1996b).
We also assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011); we assessed the following for each study.
-
Random sequence generation (checking for possible selection bias). We assessed the method used to generate the allocation sequence as: low risk of bias (any truly random process, eg random number table; computer random number generator); unclear risk of bias (method used to generate sequence was not clearly stated). We excluded studies using a non‐random process that were therefore at high risk of bias (eg odd or even date of birth; hospital or clinic record number).
-
Allocation concealment (checking for possible selection bias). The method used to conceal allocation to interventions before assignment determines whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment. We assessed the methods as: low risk of bias (eg telephone or central randomisation; consecutively numbered sealed opaque envelopes); unclear risk of bias (method was not clearly stated). We excluded studies that did not conceal allocation and were therefore at high risk of bias (eg open list).
-
Blinding of outcome assessment (checking for possible detection bias). We assessed the methods used to blind study participants and outcome assessors from knowledge of which intervention a participant received. We assessed the methods as: low risk of bias (study stated that it was blinded and described the method used to achieve blinding, eg identical tubes containing gel, or identical plasters; matched in appearance and smell); unclear risk of bias (study stated that it was blinded but did not provide an adequate description of how blinding was achieved). We excluded studies that were not double‐blind and therefore at high risk of bias.
-
Incomplete outcome data (checking for possible attrition bias due to the amount, nature, and handling of incomplete outcome data). We assessed the methods used to deal with incomplete data as: low risk of bias (less than 10% of participants did not complete the study or used 'baseline observation carried forward' analysis, or both); unclear risk of bias (used 'last observation carried forward' analysis and 10% of participants or more did not complete the study); or high risk of bias (used 'completer' analysis and 10% of participants or more did not complete the study).
-
Size (checking for possible biases confounded by small size). Small studies overestimate treatment effects, probably due to methodological weaknesses (Dechartres 2013; Nüesch 2010). We assessed studies as at low risk of bias if they had at least 200 participants in each treatment arm, at unclear risk if they had 50 to 199 participants, and at high risk if they had fewer than 50 participants.
Measures of treatment effect
We planned to estimate treatment effect for dichotomous outcomes by calculating risk ratio (RR) to establish statistical difference, and numbers needed to treat to benefit (NNT) and pooled percentages as absolute measures of benefit or harm (Cook 1995; Morris 1995). We did not plan to pool continuous outcome data because response to treatment does not usually follow a normal (Gaussian) distribution.
However, because of clinical and methodological heterogeneity between the studies, we did not undertake any pooled analysis.
Unit of analysis issues
Randomisation was to the individual participant.
Dealing with missing data
Analysis was conducted according to the intention‐to‐treat (ITT) principle. Missing data and imputation methods were incorporated in the risk of bias assessment.
Assessment of heterogeneity
We planned to examine the heterogeneity of studies with regard to intervention effects visually using L'Abbé plots (L'Abbé 1987). However, we performed no pooled analysis.
Assessment of reporting biases
We did not formally assess reporting biases.
Data synthesis
We planned to pool data where the same or comparable measures were used. We did not carry out any data synthesis because of substantial clinical and methodological heterogeneity in the included studies.
Subgroup analysis and investigation of heterogeneity
We planned subgroup analyses for different painful conditions, but we performed no pooled analysis.
Sensitivity analysis
We planned sensitivity analyses for different treatment regimens and for participants with different baseline levels of 25‐hydroxyvitamin D (the effects of vitamin D treatment might differ in people who are deficient or insufficient in 25‐hydroxyvitamin D versus people who have sufficient levels), but we performed no pooled analysis.
Results
Description of studies
Results of the search
The electronic database searches in CENTRAL (Appendix 1), MEDLINE (Appendix 2), and EMBASE (Appendix 3) conducted for this update identified 792 hits after de‐duplication, of which 769 were excluded based on titles and abstracts. We examined 23 articles fully, and included four studies (McAlindon 2013; Sanghi 2013a; Schreuder 2012; Wepner 2014). Other searches identified four studies to be examined as full texts, of which we included two (Hansen 2014; Salesi 2012) (Figure 1).
Study flow diagram. Six NEW studies
In addition, we found 12 studies ongoing (ACTRN12610000322033; Cao 2012; CTRI/2014/06/004658; EUCTR 2009‐015835‐34‐IT; EUCTR 2014‐000047‐33‐NL; ISRCTN94818153; NCT00279461; NCT01023490; NCT01351805; NCT01426347; NCT02002000; NCT02243800). Details are in the Characteristics of ongoing studies table. We were unable to determine exactly how many participants were enrolled in these studies.
Included studies
For this update, we included six new studies with 517 participants (Hansen 2014; McAlindon 2013; Salesi 2012; Sanghi 2013a; Schreuder 2012; Wepner 2014), and four studies with 294 participants from the earlier review (Brohult 1973; Di Munno 1989; Warner 2008; Yamauchi 1989).
The 10 included studies assessed people with rheumatoid arthritis (Brohult 1973; Hansen 2014; Salesi 2012; Yamauchi 1989), knee osteoarthritis (McAlindon 2013; Sanghi 2013a), polymyalgia rheumatica (Di Munno 1989), 'non‐specific' musculoskeletal pain (Schreuder 2012), 'diffuse' musculoskeletal pain (Warner 2008), and fibromyalgia (Wepner 2014). The mean age ranged from 42 to 68 years, and most studies included more women than men (range of women 41% to 100%).
Vitamin D preparation, dose, and duration varied considerably between the studies. Yamauchi 1989 used 1‐hydroxycholecalciferol, Di Munno 1989 and Salesi 2012 used 25‐hydroxyvitamin D, the other studies used calciferol preparations. Wepner 2014 gave some people in the active treatment group as little as 1200 IU of calciferol daily for 25 weeks, Brohult 1973 gave 100,000 IU of calciferol daily for one year. McAlindon 2013 used dose adjustments to keep vitamin D levels in a target range and treated over two years. Schreuder 2012 gave a single dose of 150,000 IU vitamin D.
Not all studies reported concentrations of vitamin D or metabolites in blood at the start and end of the treatment period. For those studies that did, results for placebo and treatment are in Appendix 4, and in Figure 2 and Figure 3. This shows that most studies demonstrated major change with treatment but not placebo, but only two had initial serum 25‐hydroxyvitamin D levels below 20 ng/mL signifying pronounced deficiency (Schreuder 2012; Warner 2008), one had initial levels of about 40 ng/mL (Salesi 2012), and that others had levels in the insufficiency range (Hansen 2014; McAlindon 2013; Wepner 2014). Weighted mean levels with placebo rose by 3 ng/mL and with active treatment by 14 ng/mL.
Vitamin D concentrations at start and end of treatment with placebo. The size of the symbols is proportional to the number of participants (inset scale).
Vitamin D concentrations at start and end of treatment with vitamin D supplementation (active treatment). The size of the symbols is proportional to the number of participants (inset scale).
Full details are in the Characteristics of included studies table.
Excluded studies
We excluded 21 studies from records identified in searching for this update; 11 studies were excluded in the original review (Abou‐Raya 2014; Arden 2006; Arden 2014; Björkman 2008; Chlebowski 2013; de Nijs 2006; Ding 2013; Dottori 1982; Felson 2013; Gendelman 2014; Grove 1981; Kessenich 2010; Knutsen 2014; Kragstrup 2011; Krocker 2008; Lyritis 1994; Mascitelli 2012; McAlindon 2010; Osunkwo 2012a; Osunkwo 2012b; Osunkwo 2012c; Ringe 2000; Ringe 2004; Ringe 2005; Ringe 2007; Sakalli 2012; Sanghi 2011; Sanghi 2012; Sanghi 2013b; Sanghi 2014; Wicherts 2011; Wynn 2013). Exclusions were largely because studies were reported as short conference abstracts rather than full publications, were not double‐blind randomised controlled trials, or because it was not clear that all participants had a chronic painful condition. Full details are in the Characteristics of excluded studies table.
Risk of bias in included studies
Five studies scored the maximum of 5/5 on the Oxford Quality Scale (McAlindon 2013; Sanghi 2013a; Schreuder 2012; Warner 2008; Wepner 2014), one scored 4/5 (Hansen 2014), three scored 3/5 (Brohult 1973; Salesi 2012; Yamauchi 1989), and one scored 2/5 (Di Munno 1989). We have reported individual scores in the Characteristics of included studies table. Points were lost mainly due to failure to describe methods of randomisation and blinding adequately.
We assessed the risk of bias in included studies using the 'Risk of bias' tool, and have summarised the findings below and in Figure 4. Details for each study are in the Characteristics of included studies table.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
Allocation
All the studies stated that they were randomised, but four did not provide adequate descriptions of the methods used to generate the random sequence or conceal the random allocation (Brohult 1973; Di Munno 1989; Salesi 2012; Yamauchi 1989).
Blinding
All the studies blinded participants and study personal, but two did not provide adequate descriptions of the methods used to blind the interventions (Brohult 1973; Di Munno 1989).
Incomplete outcome data
Warner 2008, Wepner 2014, and Yamauchi 1989 were at high risk of bias due to incomplete outcome data. The other studies were at unclear or low risk.
Selective reporting
While there was no clear evidence of selective reporting, only six of the 10 studies reported on vitamin D levels at the beginning and end of the trial (Hansen 2014; McAlindon 2013; Salesi 2012; Sanghi 2013a; Warner 2008; Wepner 2014). The others did not report either time point (Brohult 1973; Di Munno 1989; Yamauchi 1989), or only baseline vitamin D levels (Schreuder 2012).
Other potential sources of bias
None of the included studies enrolled 200 or more participants per treatment arm, which we consider is the minimum required to give confidence in the results. Six studies enrolled fewer than 50 participants in each treatment arm and we judged them to be at high risk of bias (Brohult 1973; Di Munno 1989; Hansen 2014; Schreuder 2012; Warner 2008; Wepner 2014). The remaining four studies enrolled between 50 and 199 participants per treatment arm and we judged them to be at unclear risk of bias (McAlindon 2013; Salesi 2012; Sanghi 2013a; Yamauchi 1989).
Effects of interventions
Because the studies were heterogeneous with regard to the included participants, interventions (amount and schedule of administration of vitamin D, co‐interventions), duration, and outcomes reported, we did not undertake a pooled analysis. Below, we discussed the studies by type of outcomes (efficacy or safety) and by the conditions treated.
Efficacy outcomes
Appendix 4 summarises the efficacy outcomes in individual studies.
Rheumatoid arthritis
Four studies investigated participants with rheumatoid arthritis (Brohult 1973; Hansen 2014; Salesi 2012; Yamauchi 1989).
In a trial of calciferol 100,000 IU daily conducted over 12 months, Brohult 1973 showed that the consumption of analgesics and anti‐inflammatory drugs decreased significantly in the calciferol group while there was no change in the placebo group (no direct between groups comparison was reported). Hansen 2014 investigated ergocalciferal 50,000 IU plus calcium initially three times weekly and then twice monthly for a total duration of 12 months, finding a worse outcome with vitamin D than with placebo (just calcium) in one pain comparison; the Patient Assessment of Global Health was also worse with vitamin D. Salesi 2012 gave participants vitamin D 50,000 IU weekly plus methotrexate or methotrexate alone for a duration of 12 weeks. There was no difference between the groups in visual analogue scale pain scores. None of these three studies investigated our preferred responder pain outcomes. One trial of alfacalcidol (1‐hydroxycholecalciferol) over 16 weeks found that similar proportions of participants reported "very much" or "much" improvement in the Patient Global Impression of Change in the two treatment groups (alfacalcidol 1 to 2 μg) and in the placebo group (Yamauchi 1989).
Overall, the studies in rheumatoid arthritis did not show a consistent direction of effect of vitamin D. Hansen 2014 and Salesi 2012 reported baseline and end‐of‐trial vitamin D levels, finding greater increases in vitamin D levels with vitamin D than placebo, but the increases were small to moderate on average. Only in Hansen 2014 did participants have vitamin D levels in the insufficiency range at the start of the trial and reach sufficiency (30 ng/mL) with vitamin D treatment. In Salesi 2012, the participants were already vitamin D sufficient at the start of the treatment.
Knee osteoarthritis
Two studies investigated vitamin D for knee osteoarthritis (McAlindon 2013; Sanghi 2013a).
McAlindon 2013 initially administered cholecalciferol 2000 IU daily with later dose adjustments, or placebo. The active treatment moved participants from vitamin D insufficiency at the beginning of the trial well into the sufficiency range at trial end, after two years. Pain responder outcomes were not reported. Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain and function score changes were not significantly different between the groups. Sanghi 2013a administered vitamin D 60,000 IU daily for 10 days followed by the same dose given once a month, or placebo, for 12 months. This moved participants in the active treatment group from vitamin D deficiency at trial beginning to sufficiency at trial end. Although Sanghi 2013a did not report any responder pain outcomes, the comparison of average visual analogue pain scale changes revealed a significant difference in favour of the vitamin D treatment. WOMAC pain score changes and total score changes also indicated superiority of vitamin D over placebo in that study.
Other painful conditions
One trial of 25‐hydroxyvitamin D (35 μg daily for 25 days a month) for nine months in people with polymyalgia rheumatica found that subjective pain on movement decreased in both vitamin D and placebo groups over time, with little difference between the groups (no statistics for between‐group comparisons reported) (Di Munno 1989).
Schreuder 2012 investigated non‐specific musculoskeletal pain in non‐Western immigrants who were vitamin D deficient at the onset of the study, and used a one‐time treatment of vitamin D 150,000 IU or placebo. After six weeks, the combined outcome of "much less pain" and "less pain" revealed a significant difference in favour of vitamin D. End‐of‐trial vitamin D levels were not reported.
One study of ergocalciferol 50,000 IU for the treatment of diffuse musculoskeletal pain over three months investigated pain with a visual analogue scale (finding no significant difference between vitamin D and placebo groups) and with a functional pain score (with an outcome that slightly favoured placebo in one comparison) (Warner 2008) (Appendix 4). In Warner 2008, participants were, on average, vitamin D deficient at trial onset and at the lower end of the sufficiency range at trial end in the active treatment group. Although vitamin D levels also increased in the placebo group, people in the placebo group on average remained in the deficiency range at trial end.
Wepner 2014 treated participants with fibromyalgia with cholecalciferol 1200 to 2400 IU daily (dose depending on vitamin D levels) or placebo. Both active and placebo groups were on the borderline of the vitamin D deficiency and insufficiency ranges (around 20 ng/mL), and both groups were in the sufficiency range after 25 weeks of treatment. Although responder pain outcomes were not reported, some other pain and quality of life outcomes suggested benefit with vitamin D.
Safety outcomes
Appendix 5 summarises the adverse event and withdrawal outcomes in the individual studies.
Overall, adverse events were infrequent and, where reported, the numbers of participants with any adverse event or with any serious adverse events did not seem to differ systematically between vitamin D and placebo groups. Salesi 2012 found 7/60 participants had adverse events with methotrexate plus vitamin D versus 3/57 with methotrexate alone.
All‐cause withdrawals were reported in all but two studies, with no consistent difference between vitamin D and placebo. Lack of efficacy withdrawals and adverse event withdrawals, when reported, also did not show a consistent difference between active treatment and placebo.
Wepner 2014 reported that one person receiving vitamin D had mild hypercalcaemia (2.71 mmol/L); the study medication was interrupted and the serum calcium level returned to the normal range. Brohult 1973 reported that after 10 months of treatment, one participant in the calciferol group had a serum calcium level (3.5 mmol/L) that exceeded the upper limit of the normal range, and that two participants in the calciferol group reported polydipsia and increased frequency of urination on one occasion (these can be symptoms of hypercalcaemia, but they are not specific for it). None of the other trials reported cases with hypercalcaemia as an adverse event.
Discussion
Summary of main results
This updated review found no convincing beneficial effect of vitamin D supplementation on chronic pain. The review has more information than a previous version of the review that reached similar conclusions.
This result has to be interpreted with caution. A major concern must be whether the studies were sufficiently sensitive to detect a difference with treatment. Only one study demonstrated unequivocal vitamin D deficiency before treatment, with 25‐hydroxyvitamin D levels below 10 ng/mL (Schreuder 2012). Other studies had initial values generally of 20 ng/mL or higher. The effects of vitamin D supplementation on 25‐hydroxyvitamin D levels was modest; although the five studies reporting end‐of‐study levels reported average 25‐hydroxyvitamin D levels above 30 ng/mL, two were borderline, and the weighted mean increase of 14 ng/mL was only a 50% increase on the initial level of 28 ng/mL. Some trial participants were vitamin D sufficient at the onset of the trial (Salesi 2012), thus, a vitamin D effect might be difficult to demonstrate in any case. Therefore, there would appear to be major questions as to whether the studies had designs able to sensitively demonstrate effects of vitamin D on pain.
Another reason why results have to be interpreted with caution is that responder outcomes of clinically useful benefit were reported only infrequently (Dworkin 2008; Moore 2010). Relying on group mean values for pain intensity or pain relief is problematic because the underlying distributions of individual data are often skewed, and using average data from such skewed distributions can produce misleading results (McQuay 1996; Moore 2013). This is why dichotomous responder analysis has been suggested to produce more useful results in clinical practice both in the acute (Moore 1998), and chronic (Moore 2008), pain settings.
With these caveats in mind, we found that there was no consistent effect of vitamin D on pain in rheumatoid arthritis. In knee osteoarthritis, one (smaller and shorter) study reported significantly better outcomes with vitamin D (Sanghi 2013a), while another (larger and longer) study found no difference between the groups (McAlindon 2013). With other painful conditions there is some suggestion of benefit in fibromyalgia (Warner 2008), and non‐specific musculoskeletal pain in non‐Western immigrants (Schreuder 2012), while there was no evidence of superiority of vitamin D over placebo in one study on polymyalgia rheumatica (Di Munno 1989). One study on diffuse musculoskeletal pain even reported benefit with placebo in one outcome (Warner 2008). Overall, there is no consistent pattern that vitamin D treatment was associated with greater efficacy than placebo in chronic painful conditions. Adverse events and withdrawals were comparatively infrequent, also with no consistent difference between vitamin D and placebo.
Overall completeness and applicability of evidence
A number of potentially eligible studies had methodological limitations and had to be excluded from this review. Other potentially eligible studies were identified by searching trial registries but were not (yet) published and did not have data available for analysis. The 10 studies included in this review were conducted at different times; in different locations; treating a number of painful conditions in participants with vitamin D levels in the deficiency, insufficiency, and sufficiency ranges with different vitamin D doses and schedules of administration over different periods of time; and assessing different outcomes. This heterogeneity meant that no meaningful pooled analysis could be undertaken.
Because of these limitations and the infrequent use of the responder pain outcomes now thought to be most informative of clinical benefit, the evidence has to be judged to be incomplete. Based on the included studies, there is no clearly demonstrable benefit associated with vitamin D in chronic pain in adults. It can be argued that if substantial benefit was present across different chronic painful conditions, it would have become apparent; however, this was not the case.
Quality of the evidence
Though the included studies were largely assessed to be at low or unclear risk of bias, the relative paucity of trials and participants in high quality studies, together with the inadequacy of outcome measures used, implies that the evidence base regarding the use of vitamin D in chronic painful conditions is weak.
Potential biases in the review process
Because of adherence to quality criteria and an extensive search strategy, biases in the review process were minimised.
Agreements and disagreements with other studies or reviews
This review agrees with a previous review on the subject in that there is little high quality evidence for an effect of vitamin D in chronic pain, despite the existence of a number of studies of lower methodological quality (not double‐blind randomised controlled trials) that suggest an effect (Straube 2009). As we have previously stated, there is a 'beguiling attraction' in a link between low 25‐hydroxyvitamin D levels and chronic pain (Straube 2009). If effective, vitamin D would be a simple and inexpensive treatment with limited adverse effects. However, good evidence for this effectiveness is lacking.
There are a number of examples where early observational studies and treatment studies of lower methodological quality appear to demonstrate effects that are not subsequently confirmed by high quality trials. For example, the use of hyperbaric oxygen in multiple sclerosis appeared beneficial based on observational and non‐randomised studies, but subsequent randomised controlled trails found no benefit (Bennett 2001). Likewise, for the use of transcutaneous electrical nerve stimulation for postoperative pain, non‐randomised trials were largely positive while randomised trials were largely negative (Carroll 1996). Our previous review on vitamin D for chronic pain came to a similar conclusion when comparing high and low quality treatment studies (Straube 2009). This present review, updated to include 10 studies, relied on a wider search and more strict inclusion criteria than our previous review (Straube 2009), but also did not identify consistent evidence of an effect of vitamin D.
Study flow diagram. Six NEW studies
Vitamin D concentrations at start and end of treatment with placebo. The size of the symbols is proportional to the number of participants (inset scale).
Vitamin D concentrations at start and end of treatment with vitamin D supplementation (active treatment). The size of the symbols is proportional to the number of participants (inset scale).
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
